modern_hopen_hole log interpretation.pdf

MODERN

INTERPRETATION OPEN-HOLE LOG

c3552949

Z -- 111111 Q

i-.I O

Z

2

3 Lu Q

DEDICATION

Copyright (c) 1983 by PennWrli Publishing Company 1421 South Sheridan RoadiP.0. Box 1260 Tulsa, Oklahoma 74101

LiDrurg of Congress Caiaioging ! t i Yuhlicotioii Dato

Dewan, John T. Modcrri open-hole log intrrpretaticiri

All right5 rr5er\rd. N o part of this book nia) be reproduced, stored in a retrieval systcvn, or transcribed in ariy form or by ariy means, electronic or mechanical, including photocopying and recording, without the prior written permission of the publisher.

Printed i n the United States of America

This book is dedicated to my former colleagues at Schlumberger, who did much to advance the science of well logging, and to my present associates and my family, who patiently endured its preparation.

3 4 5 87 86 85

i v

CONTENTS

DEDICATION v

INTRODUCTION xi

1 The Logging Environment 1 The borehole 1 Logging procedure 2 The undisturbed reservoir 4 Disturbance caused by drilling 8 Summary 13 References 15

2 Evaluation of Hydrocarbons 17 Fundamental interpretation relations 17 The basic interpretation procedure 22 Impact of invasion on resistivity measurements 26 Summary 31 References 33

3 Permeable Zone Logs 35 Spontaneous Potential (SP) Log 35

Source of spontaneous potential 36 SP behavior over a long log 39 Shape of the SP curve 40 Computation of R, from the SP 43 SP log in shaly sands 49 SP anomalies due to vertical migration of filtrate 49 SP anomalies due to noise 50

The Gamma Ray (GR) Log 50 Basic GR logs 50 Spectral GR logs 53 Statist ica I f Iuc t uat ions 57 Summary 60 References 61

v i i i CONTENTS CONTENTS i x

4 Resistivity Logs 63 Classification and Application 63

Electrical survey (ES) tools 65 Fresh mud tools 66 Salt mud tools 68

Ranges of application of induction logs and Laterologs 69

The spherically focused log (SFL) 76 Log presentation 77

Dual Laterolog-R,, Logs 82

Dual Induction-Spherically Focused Logs 70

The DLL-MSFL tool 83 Depth of investigation 86

Characteristics of the MicroSFL (MSFL) 88 Log presentation 88 Quick-look hydrocarbon indication 89 Summary 93 References 94

5 Porosity Logs 95 The current trend in poroslty logging 95 Recent developments 96

Compensated Density and Litho-Density Logs 97 The compensated density tool 98 Porosity derivation from the density log The Litho-Density log 108

Lithology interpretation with pb-P, curves 114

Neutron tool evoiution 117 The Compensated Neutron 117 Combined Density-Neutron interpretation 128

The Dual Porosity Compensated Neutron log 136 Compensated Sonic and Long-Spaced Sonic Logs 139

The borehole compensated log 141

Porosity determination from Sonic logs 146 The Long-Spacing Sonic log 158

105

Compensated Neutron and Dual Porosity Neutron Logs 115

Electromagnetic Propagation-Microlog Combination 170 The EPT-ML sensor array 172 The Microlog 173 The Electromagnetic Propagation Log 177 Summary 191 References 194

6 Clean Formation interpretation 199 Resistivity-Porosity Crossplots 199

The Hingle plot 200 The Pickett plot 207 Range of uncertainty in calculated water saturations 209

The M-N plot 21 1 The MID plot 214 The Litho-Density-Neutron method 217 Trends in multimineral identification 223 References 225

Multimineral Identification 211

7 Shaly Formation Interpretation 227 The nature of shale 230 Shale or clay distribution in shaly sands 231 Shaiy sand interpretation models 236 Cation exchange capacity 237 Shale porosity and conductivity 243 Application of the dual-water method to shaiy sands 247 Summary of dual water interpretation 257 Summary of earlier shaly sand interpretation methods 261 References 265

8 Prediction of Producibility 267 Flow relations 267 Absolute. relative, and effective permeabilities 269 irreducible water saturation 271 Estimation of permeability from logs 279

x i ¡ INTRODUCTION Chapter 1

'I'lie third and current plinse, which bcgan abont 1'370, ma!' be called t h r log procrssing era. \I'ith the advent of computers, it has beconie possible to analyze i n much greater detail the \vealth of data sent uphole by the logging tools. Log processing centers, providing sophisticated interpretation o1 digitized logs transmitted b y telephone and satellite, haire been set up b!, service companies in strategic locatioiis. Logging trucks have been fitted n,ith computers that perinit computation of quick-look l o g s at the wellsite. At thesanic time logging tools have been coiribined to tlie p i n t that a full set of logs can be obtained on a sirigle run.

Tlic present state of the art is that logs are adeqiiiite to determine hydrocarbons in situ in inediuni- to high-porosity formations but are piislied to their limits in low-porosit!., slial!,, rnixed-litliolog!~ situations. .\lore precise deterniination of the niatris makeup, including ainounts and t ! p s of c l a ~ . present, is needed. Promising developrnt~its are uiider\vay.

Ad\.ances are being made i n predicting the proclucibilit!~ of liydrocarboiis found in place. but tlie critical factor: a contiriuoiis perriieabilit!. log, is still lacking. Meanwhile, point-hy-point permeability and pressure \zaliies can be 01)tairied b!. repeat formation testirig. a tecliiiiqut. that i s finding increased use.

Developments i n the testing stage promise to provide more precise lithology information. better movable oil determination, and additioiinl rnechaiiical properties of formations. Uriciiiestioria~)l!.. ansuws obtairial>le froin logs will continue to become more accurate and broader in scope.

THE LOGGING ENVIRONMENT eiotively little is learned about the producing potential of a well as i t is being drilled. This is a surprise to the uninitiated, who have visions of early gushers. But the drilling mud actually pushes hydrocarbons,

if encountered, out of the way and prevents their return to the surface. Examination of returned cuttings indicates the general lithology being penetrated and may reveal traces of hydrocarbons, but it allows no estimates of the amount of oil or gas in place.

Well logs furnish the data necessary for quantitative evaluation of hydrocarbons in situ. Modern curves provide a wealth of information on both the rock and fluid properties of the formations penetrated. From the point of view of decision-making, logging is the most important part of the drilling and completion process. Obtaining accurate and complete log data is imperative. Logging costs account for only about 5 % of completed well costs, so it is false econoniy to cut corners in this phase.

THE BOREHOLE

MJhcn the logging engineer arrives at the wellsite with his highly instru- mented logging unit, he finds readj. to be surveyed a borehole that has the following: characteristics:

an average depth of about 6,000 f t but which may be anywhere between 1,000 and 20,000 ft an average diameter of about 9 in. but which can be between 5 in. and 15 in. a deviation from vertical that is usually only a few degrees on land but typically 20-40" offshore a bottom-hole temperature that averages about 150°F but may be between 100°F and 350°F a mud salinity averaging about 10,000 parts per million (ppm) but which can vary between 3,000 and 200,000 ppm; occasionally the mud may be oil based a mud weight averaging about 11 lbigal but which can vary from 9 to 16 lb/gal a bottom-hole pressure averaging perhaps 3,000 psi but which can be as low as 500 and as high as 15,000 psi

4

,

References 344

9 Wellsite Computed Logs 317

Hard-rock, salt-mud logging suites 343 Special situations 346 References 349

index 351

INTRODUCTION

The aim of this book is to present modern log interpretation as simply and concisely as possible. The book is written for the geologist, petrophysicist, reservoir engineer, or production engineer who is familiar

him with computer-processed logs generated by the service companies at the wellsite.

Accordingly, obsolete logging tools are mentioned only in perspective. Very brief descriptions of the instruments in common use indicate how they apply to different logging conditions. Salient features of new tools, including Spectral Gamma Ray, Litho-Density, Dual Porosity Neutron, and Long-Spacing Sonic, emphasize how these tools fit into the everyday

concerned how a tool

logs i s run in a liquid-filled open hole to locate hydrocarbons in place and where promising zones are then tested to evaluate their producibility. Abnormal situations such as empty hole, water well, and geothermal and mineral logging are not included.

To provide a little perspective, well logging is in its third major development stage. The first 20 years, from 1925-1945, saw the introduction and gradual worldwide acceptance of the so-called ES (Electrical Survey) logs. These logs were run with simple downhoie tools and, while quite repeatable, were often difficult to interpret.

The second phase, from 1945-1970, was a major tool development era, made possible by the advent of electronics suitable for downhole use. Focused electrical devices were introduced, having good bed resolution and various depths of penetration. A variety of acoustic and nuclear tools were developed to provide porosity and lithology information. There was a progression through second- and even third-generation tools of increasing capability and accuracy. Simultaneously, much laboratory and theoretical work was done to place log interpretation on a sound, though largely empirical, basis.

x i

2 ESSENTIALS OF MODERN OPEN-HOLE LOG INTERPRETATION

a sheath of mud cake on all permeable formations that averages about 0.5 in. in thickness but may be as little as 0.1 in. and as much as 1 in. an invaded zone extending a few inches to a few feet from the borehole in which much of the original pore fluid has been displaced by drilling fluids

Even more severe conditions are occasionally encountered. In any case it is a challenging environment from which to derive accurate information about the state of the formations as they were prior to any drilling disturbance.

LOGGING PROCEDURE

Accustomed to the challenge, the logging crew proceeds to align the truck with the well, spool the logging cable through the lower and upper sheave wheels, and connect the logging tools. The engineer performs the surface checks and calibrations. After this, the logging array is dropped to bottom as quickly as practicable. Once on bottom the downhole calibrations are carried out, recording scales are set up, and the crew “comes up logging” (Fig. 1-1). Survey speed is maintained constant, between 1,800 and 5,400 ftlhr, depending on the logging tools.

‘The logging string is typically 3% in, in diameter and 20-50 ft long. It usually consists of several different tools in tandem. The most important is the tool that measures the electrical resistance of the formation because increased resistance occurs when water is replaced by hydrocarbons. Accompanying the resistivity tool is at least one tool that measures porosity and one that distinguishes permeable from nonpermeable zones. The basic logs may be obtained on a single run in the hole or may require two runs with different logging tools. Operating power for the tools is sent down one pair of insulated conductors inside the armored logging cable, and logging data are transmitted to the surface on the remaining five conductors.

In recent years the major service companies have been replacing older surface instrumentation with completely computer-controlled systems that are much more versatile and easier for the engineer to operate (Fig. 1-2). * Logging data are digitized and fed into the computer where they are

‘Denoted Cyber Service IJnit (CSU) by Schlumberger, Computer Logging System (CIS) by Dresser Atlas, Digital Logging System (DLS) by Welex, and Direct Digital Logging (DDL) by Gearhart.

I THE LOGGING ENVIRONMENT 3

processed and output to paper, film, and magnetic tape recorders. The engineer controls the system almost entirely with commands froin the key- board. At the same time, he monitors the output on a screen wliich, during

_-_____L__-

Fig 1-1 Wellsite setup for logging [courtesy Gearhart)

f

Fig. 1-2 Computerized surface instrumentation (courtesy Schlumberger)

logging, continuously displays the last 100 ft logged. The options in cali- brating, depth shifting, averaging, computing, and scaling the logs are virtually unlimited with the computer. Further, after the logging is completed, the taped information can be played back and the various logs edited, combined, and run through interpretation programs to provide fully interpreted logs at the wellsite.

THE UNDISTURBED RESERVOIR

An idealized view of porous hydrocarbon-bearing reservoir rock is shown in Fig. 1-3. The rock matrix consists of grains of sand, limestone,

T 5

dolomite, or mixtures of these. Between the grains is pore space filled with water, oil, and perhaps gas. The water exists as a film around the rock grains and as hour-glass rings at grain contacts; it also occupies the very fine crevices. The water forms a continuous path, although very tortuous, through the rock structure.

& spaces.

The rock properties important in log analysis are porosity, water saturation, and permeability. The former two determine the quantity of gasor oil in place, and the latter determines the rate at which that hydrocarbon can be produced.

Porosity

011 + water Oil + gas + water

Fig. 1-3 Hydrocarbon-bearing rock

6 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION

SLJ: / O Q * / . - 5d'.

Tlie fraction of pore space containing water is termed water safurafion, denoted S,, . The remaining fraction Containing oil or gas is termed hydrocarbon safuratioti, Sh, which of course, equals (1 - C,,,). The general assumption is that the reservoir was initially filled with water and that over geologic time oil or gas that formed elsewhere migrated into the porous formation, displacing water from the larger pore spaces. However. the migrating hydrocarbons never displace all of the interstitial water. There is an irreducible water sattiratioti, S,,,, representing the water retained by surface tension on grain surfaces. at grain contacts, and in the smallest interstices. Its valiie varies from about 0.05 in very coarse formations with low surface area to 0.4 or more in very fine-grained formations with high surface area. The irreducible \vater will not flow Lvhen the formation is put on production.

The fraction of total formation \.oliiine that is hydrocarbons is then 4 Sh or I$ (1 - S,v), A major objective of logging is to determine this qiiantity, It can vary from zero to a maximum of 4 (1 - S,,,,).

Permeability

Permeability. denoted k, i s the Elowability of the formation. It i s a measure of the rate at which fluid will flow through a given area of porous rock under a specified pressure gradient. It is expressed in millidarcies (md); 1,000 riid is a high value and 1.0 md is a low value for producing formations.

In contrast to porosity, permeability depends strongly on absolute grain size of the rock. Large-grained sediments with large pores have high permeabilities, whereas fine-grained rocks with small pores and more tortuous flow paths ha\Te l o ~ v permeabilities.

Table 1-1 lists porosities and permeabilitiesof some well-known producing forniations. Porosity varies only by a factor of 3, whereas permeability varies by a factor of about 4,000. We can infer that the\Voodbine formation with extremely high permeability is exceptionally coarse sand, whereas the Ctrawn formation of the same porosity but low permeability is a very fine- grained sandstone.

Hydrocarbon-Bearing Rocks

Hydrocarbon-bearing rocks are primarily sands (SiOp,). limestones (Caco3) , and dolomites (CaCO, . MgCO,). hlost sands are transported by

TABLE 1 - 1 POROSITIES AND PERMEABILITIES OF SELECTED OIL SANDS

Sand Poroslty Permeabllity

(%I ímd)

Clinch, Lee Co. VA Wllcox, Okla. Co. OK Cut Bank, Glacier. Co. MT Bartlesville, Anderson Co. KS Olympic. Hughes Co. OK Woodbine, Tyler Co. TX Strown, Cooke Co. TX Nugget. Fremont Co. WY O'Hern, Duval Co. TX Eutaw, Choctaw Co. AL -

9.6 12.0 15.4 17.5 20.5 22 22 25 28 30 -

0.9 1 O 0 111 25 35

3,390 81

147 130 1 O0

and laid down from moving water. Tlie greater the water velocity (the energy of the environment), the coarser the sand will be. Because of this mechanism, sands tend to have fairly uniform intergranular-type porosity.

Limestones, on the other hand, are not transported as grains but are laid down by deposition from seawater. Some is precipitation from solution; some is the accumulated remains of marine shell organisms. Original pore space is often altered by subsequent redissolution of some of the solid matter. Therefore, porosity tends to be less uniform than in sands, with vugs and fissures, termed secondary porosity, interspersed with the primary porosity.

Ilolomites are formed when magnesium-rich water circulates through limestones, replacing some of the calcium by magnesium. This process generally results in a reduction of the matrix volume. Therefore, dolomitization i s an important mechanism in providing pore space for hydrocarbon accumulation.

Formations containing only sands or carbonates are called clean formations. They are relatively easy to interpret with modern logs. When such formations contain clay, they are called dirty or shaly formations. Such reservoir rocks can be quite difficult to interpret.

Clay and Shale

Clays are common components of sedimentary rock. They are alumino- silicates of the general composition AI,O3. S O 2 . (OH),. Depending on the environment in which they are formed, they may be of several basic types: rnontmorillonitr, illite, chlorite, or kaolinite.

Clays have very small particle sizes-1 to 3 orders of magnitude less than those of sand grains. Surface-to-volume ratios are very high, 100-10,000 times those of sands. Thus, clays can effectively bind large quantities of water that will not flow but that do contribute to log response.

Shales are primarily mixtures of clay and silt (fine silica) laid down from

hand, sands or carbonates containing modest amounts of clay or shale may be important hydrocarbon producers.

Accounting for clay and shale when analyzing hydrocarbon-bearing formations substantially complicates log interpretation. Consequently, in chapters 2 through 6 we establish the principles of log interpretation for clean formations and in chapter 7 take up the analysis of shaly format

During the drilling process formations may erode or cave to diameters larger than bit size, drilling fluid may invade permeable zones, and mud cakes may build up on the same zones. invasion in particular causes logging problems.

The Process of Invasion

The process of invasion is also illustrated in Fig. 1-4. Duringdriiling the mud pressure in the annulus, P,, must be kept greater than the hydrostatic pressure of fluid in the formation pores, P,, to prevent a well blowout. The differential pressure, P, - P,, which is typically a few hundred psi, forces drilling fluid into the formation. As this happens solid particles in the dríiling mud plate out on the formation wall and form a mud cake. Liquid that filters through this mud cake-the mudfiltrate-passes into the formation and pushes back some of the reservoir fluid there. An invaded zone is formed adjacent to the borehole. Fig. 1-4 Invasion process

I O ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION

Invasion involves a mud spurt, dynamic filtration, and static filtration. As the bit penetrates a permeable bed, there is an initial spurt of drilling fluid into the freshly exposed rock. This lasts for a matter of seconds until small solid particles in the downcoming mud stream (or formed by grinding action of the bit) bridge the pore entrances in the rock. Bridging is most rapid when the particle size distribution in the mud is well matched to the pore entrance distribution in the rock.

As the bit passes on, rnud cake begins to build on the newly formed borehole wall. Invasion is rapid in the beginning but slows quickly as the mud cake thickens and increases its resistance to flow. If conditions were static, the mud cake would continue to build indefinitely with filtration rate decreasing in accordance with l/&, where t is the time following spurt.

During drilling, however, flow of mud and cuttings plus abrasion caused by the turning and whipping of the drillstring continuously erode the mud cake arid even the formation itself. Once the formation ceases to erode, a dynamic equilibrium condition is reached where the mud cake thickness and the rate of filtration become constant.

When the drillstring is pulled to change the bit, the abrasive action i s no longer present and mud cake resumes building at the permeable zones under static filtration conditions. When the string is run in and drilling is resumed, the soft outer miid cake just formed will erode away and dynamic eqiiilib- rium once again will be reached.

Finally, when the drillstring is pulled for logging, static filtration will resume and soft mud cake will again build up. The additional buildup is often evidenced by logging tools measuring hole diameter less than bit diameter in permeable zones near bottom. Mud cake is typically in. in thickness at the time of logging.

Fig. 1-5 shows schematically the rate o f invasion, mud cake thickness, and depth of invasicin at a given permeable bed as a function of time since the bed was penetrated. The depth of invasion increases rapidly during the spurt and formation erosion periods. I.,ater it slows because of dynamic equilibrium and because the rate of increase in invasion depth, for a constant filtration rate, is inversely proportional io the invasion depth already reached.

Depth Of Invasion At Time Of Logging

The depth to which mud filtrate has penetrated a porous formation at the time of logging depends on several factors, principal of whích are the filtration characteristics of the drilling mud and the differential pressure between mud and reservoir. Static filtration rate of a niud is given as a water

I 1 THE LOGGING ENVIRONMENT

loss figure on the log heading. This is the amount of filtrate (in cc) passing through a filter paper in 30 min at 100 psi differential pressure and 76°F in a standarized API test. A typical figure i s 12 cc; 30 cc is considered poor wall building mud, and 4 cc is very good. Unfortunately, experiinents have shown that there is little correlation between static filtration characteristics at surface temperature and dynamic filtration at borehole teniperatilres. Consequentlj., it is not possible to predict invasion depth from available mud and drilling information. The analyst must infer this from logs.

One can predict, nevertheless. how depth of invasion for a given rnud relates to porosity. Once the mud cake has started to build. its permeability becomes low relative to that of the average formation so that almost all of the pressure differential (P, -- P,) is across the mud cake and little i s applied to the formation. The mud cake therefore controls filtration rate. Conse- quently, in a given time the same volume of fluid will invade different forniations, regardless of their porosities or permeabilities (iiriless permeability is belmv about 1 .O nid). This means depth of invasion will be niini- mum at high porosity where plenty of pore space is available for invading fluid and maximum at low porosity where little room is available. I t is

DeDth of invasion - Rate of invasion

Drill 4 Trip +Drill - E- Mud cake thickness * , I I I I I

I

10 1 O0 1,000 10 000 100.000

Fig. 1-5 Invusion effects

EVALUATION OF HYDROCARBONS

conducts no electricity. All conduction is then via the fluid in the pores. At depths below 2,000 ft, the water found in formation pores is generally fairly saline, which makes it quite conductive. Water-bearing formations therefore tend to have high electrical conductivity or, the equivalent, low electrical resistivity since resistivity is the reciprocal of conductivity.

clean formations with modern logs can often be performed without a single chart.

F U N D A M E N T A L INTERPRETATION RELATIONS

To establish the relations for hydrocarbon saturation arid to clarifjr the terms involved, let us conceptually construct an oil-bearing formation and measure its electrical properties as we do so.

Definition of R, I Visualize an open-top cubic tank one meter in all dimensions. It has

electrically nonconducting sides except for two opposite walls that are metal and serve as electrodes.

17

18 ESSENTIALS OF M O D E R N OPEN-HOLE LOG INTERPRETATION

First, the tank is filled with water containing about 10 % sodium chloride (NaC1) by weight to simulate an average formation water. A low- frequency alternating voltage, \’, is applied across the electrodes, and the resilltirig current I1 is measured (Fig. 2-la). ?’he ratio VU, (voltsiamperes) is R,. the resistivity of tlie formation water, in units of ohm-meters. This resistivity is an intrinsic property of the watcr and is a function of its salinity and teniperature. The higher these two variables, the more conductive the water will he and the lower its rcsistii-it)..

F i g 2 - 1 Llefinition of rewtiviiiec

EVALUATION O F H Y D R O C A R B O N S 19

Definition of Ro

Next, sand ic poured into the water-filled tank and tlie volume of water expelled is meawred. \i’hen the sand is level with the top, the result is a porous. water-bearing formation of one-meter dinieri~ions About 0.6 cti rn of \vater will have been expelled, so the porosity of the forniatiori will IF (1 - 0.6) or 0.4. Again the voltage is applied and a current I, is measured (Fig. 2-1 b). I2 will be less than I, siiice thcrc is less water to conduct elrctricity. The ratio \.‘/I2 is R,, the resistivity of tlie water-hearing forniation. It will be larger than H,, .

Formation Factor

The resistivity, I{<,. must be proportional to I{,< sincc c d y the water conducts. ‘ J ~ s

The proportionality constant F is ternied thefornzatiori juctor.

relation of the forni On general principles, formation factor must be related to porosity by a

4:) 3 Po2R,4., ~ - 0 - M o ’03.

(2.2) I; = 1/@P

tiwause whcn d, = 1 (all vzater. no matrix). R,, must equal H,<; and when 4) i- O (no pore water, solid matrix), H,, must be infinite since the rock itself is an insulator. E(1. 2.2 satisfies these conditions regardless of the value of m. xvhich is ternietl the ccmcntntinri cyiot iui t .

The value of ni reflects the tortiiocit!. of current flow through the maze of rock pores. If the port space consisted of cylindrical tiibes through an otherxvise solid matrix. current flo\v paths would be straight and ni would he 1 . O . In the case of porous formations, nieasurements have shown m to be 2.0 o11 the averagr. Avcepted relations for the range of porositit~s cncoiin- tcrcd in logging are

<-----_.. -Y- Artrobe. w.=z2.

F == i / ~ ‘ for limrstonci (2.3) (1 .4) E’ = 0.8119’ or 0.6216’ ’i for s a n d s ’

- - Definition of R,

I 1

1 1

oped by G.E. Archie of Shell and i s termed the Archie well-lo@@lg industy is

The whole 'pon this equation*

Eq. 2.8 shows that hydrocarbons in place can be evaluated if there are

Now an appreciable fraction of the pore water i.. replaced by oil, result- ing in the situation depicted in Fig. 2-le. The same voltage, V, is applied, and current I, is measured. It will be less than IL since even less water is

Water Saturation

Knowing R, and R,, water saturation, S,", the fraction of pore space containing water, can be calculated. Again on general principles there must be a relation of the form

!

1 F)'$ 5 .)-"-") L' " KT hi)"""""""""" ñnduct-

p* Lutu? 1% CSri\\+ mod: This is the basic equation of log interpretataon. It was initially devel-

2.8 in a nearby water sand (S, = 1) or from the SP log or from catalogs or water sample measurements; R, i s obtained from deep resistivity readings (Induction or Laterolog); and 4 is obtained from porosity logs (Density, Neutron, or Sonic).

1 i

R, = R,/C," Hydrocarbons in Place (2.5)

This relation can be used directly to calculate the water saturation of a hydrocarbon-bearing zone when an obvious w ater-bearing zone of thesatne porosity and having water ofthesamesalinity is nearby. An example would be a thick sand with an obvious water-oil contact in the middle.

In general there will not be a nearby water sand to give R,, so Eq. 2.6 will not apply. Replacing R, by Eq. 2.1 gives

s, =

Replacing F by Eq. 2.3 gives

where c= 1.0 for carbonates and 0.90 for sands.

formation volume factor B, a value slightly greater than unity, which takes into account the shrinkage of oil volume, principally by gas evolution, as it comes to the surface.

For gas, the number of cubic feet in situ is

G = 43,560.4(1 - S , ) . h . A (2.10)

'The general equation is S: = (a/<bm)(R,/R,). Values of a, rn, and n can differ from those indicated in specific caw. This is discussed in chapter 6.

ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION

___cs is the amount of gas at reservoir pressure and temperature. To - it to s tandard cubic feet at 14.7 psi and 6O0F, the number is miilti-

---===d 7' t h e quant i ty +a2 I -

= reservoir pressure, psi: if not kiiown i t may he taken as 0.46d

= formation temperature, OF, determinable from log information = deviation factor of the gas at formation temperature and prei-

where d is the vertical depth of the reservoir in feet

s u r e , obtainable from charts; it will be close to unity - e aniount of hydrocarbons actuall!. recoverable as a fraction of the IS ty in p lace will depend on the reservoir type. the initial hydrocarbon - f i o n , and the prodiiction mechanism. A reasonable estimate for pri- m r o d u c t i o n would be 20% for oil and 70% for gas.

- - ASIC INTERPRETATION PROCEDURE

Te(' basic logs are required for adequate formation evaulation. One is 4 to show permeable zones, one to give reiirtivit! of the undisturbed * t i o n and one to record porosity. An idealized set is shown in Fig 2-2. - e r m e a b l e zone log is in Track 1. the resistivitl- log in Track 2, arid the s t y log in Track 3 . The permeable zone log is either Spontanroiis - t - i a l or Gamma Hay, the resistivity i s either deep Indiiction or deep

-log, ar id the porosity log i s either Density, Neutron, or Sonic. C ' ,I\ en

- - ~~

~

__ i;i set of logc, the problem is to cletermine - e r e a r e the potential producing zones w much hydrocarbon (oil or gas) do they contain

- =tion of Productive Zones - h e f i r s t step is to locate the permeable zones. This is done by scanning

g in Track 1. It has a base line on the right arid occasional slvings to tlic - - ~~ ~


Permeable zone indication

I Resistivity 1 Porosity

Fig. 2-2 Idealized log set

26 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION

Reservoir pressure would be estimated as 0.46 x 7,300 or 3,360 psi and reservoir temperature as approximately 150'F. Assuming a gas deviation factor of 1 .O, the amount of gas in place in standard cubic feet would be, by Eq. 2.11

14.2 X lo6 X (3,360114.7) X (5201610) = 2,770 X IOf i

With a recovery factor of 0.70, the producible amount is estimated at 0.7 x 2,770 or 1,940 Mh4scf. A gas price of $31Mcf would yield a potential revenue of $5.8 million. Again, the well would be definitely worth completing.

IMPACT O F INVASION ON RESISTIVITY MEASUREMENTS

In the foregoing discussion the formation resistivity, R,, was assumed to be that of the undisturbed reservoir beyond any invasion. The difficulties in measuring that resistivity may be appreciated by considering the disturbance caused by invasion, as illustrated in Fig. 2-3.

Immediately behind the borehole wall is a flushed zone of diameter d, containing only mud filtrate of resistivity Rmf and residual hydrocarbon. The resistivity of that zone is denoted R,, and the water saturation is Sxo. The thickness of this zone is of the order of 6 in. but can be much more or much less. Behind the flushed zone is the transition zone of diameter d,, which may extend several feet. Beyond that is the undisturbed formation with resistivity R,, interstitial water resistivity R,", and water saturation C,.

The existence of invasion has forced the development of resistivity logging tools that measure as deeply as possible in an effort to read R, uninflu- enced by mud filtrate. However, no tool has been developed that can read deeply enough under all circumstances and still maintain good vertical resolution. Consequently, the industry is gradually standardizing on running three resistivity curves at the same time. One is a deep investigation curve, one is a medium curve, and one is a shallow curve. With three curves the reading of the deep one can be corrected for invasion effects to provide an R, value. As a side benefit the flushed zone resistivity and the diameter oí invasion can also be estimated.


Mud Filiraie Resistivity; Fresh Muds and Salt Muds

'To appreciate the difference in resistivity readings of shallow, medium, and deep curves, i t is necessary to consider the contrast between miid filtrate resisti\ity, R,,, and interstitial water resisti\.it!v. Rw. Mud filtrate resjsti\.ity is measured at the kvellsite by the logging engineer. He catches a sample of miid. preferably from the mud return line, places it in a mud filter press that forces filtrate through a filter paper and measures the resistivity of the filtrate in a resistivity-measuring cell. The value of H,,,, along with tlic temperature at tvhich it is measured are included in the log heading.

0- Resistivity of the zone

0- Resistivity of t h e water in t h e zone

A- Water saturation in t h e zone

1

Fig. 2-3 ldeuliiec! invasion profile (courtesy Schlurnberger)

IS are drilled with “fresh’ (low-salinity) muds that have

value. If it were deep, the IL,, would he very influenced by the mud filtrate and would read close to LL,. Thus, the position of the IL,,, curve relative to the others is a quick-look indicator of invasion depth. Finally, we can see that the shales have no permeability because they have not invaded; all curves read alike.

In wells drilled with saturated salt muds, the positions of the three resistivity curves in water-bearing intervals will reverse. The shallow curve will read lowest and the deep curve will read highest.

Invasion Profiles

Fig. 2-5 shows three different resistivity profiles proceeding from the flushed to the undisturbed zone for the fresh mud case where R, > R,. The assumption made in deriving a correction for R, with three resistivity curves is a step profile-an abrupt change from R,, to R, at an equivalent diameter

29

d,. In actuality the resistivity will change gradually from the R, value at diameter df to the R, value at diameter d,, as shown by the dashed curve of

with high hydrocarbon saturation wherein invading filtrate displaces hydrocarbon faster than the interstitial water. This creates an annulus or bank of formation water where the resistivity is temporarily lower than either R,, or R, (Fig. 2-5, dotted line). It is a transient phenomenon, lasting only a few days.?

Fig. 2-4 Effect of invasion on resistivity measurements (Courtesy Schlumberger)

30 ESSENTIALS O F MODERN OPEN-HOLE LOG INTERPRETATION

Movable Oil Calculation Invasion has one redeeming feature: It can provide information on

hydrocarbon prodiicibility through comparison of li!.drocarbon saturation in the flushed and undisturbed zones. The difference between tlie tivo saturations represents hydrocarbon that has been pushed back by invading fluid and therefore should be recoverable on production. at least hy a \vater drive simulating invasion in reverse.

The water saturation relations already developed can be applied to the flushed zone provided appropriate resistivity values are used for the fluid- filled rock, &,. and the pore fluid, Rmf. Following Eq. 2.8, the water saturation in the flushed zone is

(2.12)

‘The hydrocarbon saturation (1 - S,,,) will be less than that in the iininkaded zone, (1 - S,,), by virtue of the displacement cauíed by filtrate. Therefore, the movable oil saturation is the difference, which is (SXq -- S,,). By Eq5. 2.12 and 9.8

t 2. = > ._ E ?? c .- I

E 8 U

O Distance from borehole wall -

Fig 2-5 Resistivity profiles- step, gradual, and annulus

EVALUATION O F HYDROCARBONS 31

‘This is the fraction of pore fluid that constitutes moved oil. It reprcsents an upper limit of M hat might be recovered on production.

As an example, assume a limeytone formation with C$ == O . 18. R,, = 0.04, It,, , , = 0.50, H t = 10, and R,,, = 25 ohm-ni

That is. 43% of the reservoir pore space constitutes movable oil. Alterna- tively, the bulk \rolunie fraction of movable oil is d (Sxo - S,v) or 7 . 7 % .

Eq. 2.13 works well in salt-mud situations but tends to overestimate the amount of movable oil in fresh-mud conditions. Recent investigation has led to the conclusion that all of the connate water in tlie hydrocarboil-bearing zone niay not be replaced by mud filtrate; some may be shielded hy the rrGdual oil and left in place.‘ This iiicomplete replacement does not matter \vhen R,,, = R,,,. as i n tliecase of salt mud. However for fresh miid where >:> R,v. R,, is too high a value to use for flushed-zone \vater rcsistivity. The error can be large when the oil is heavy arid residual oil satiiration is high. ?’lie same overestimation occurs if the invasion is so shallow that the R,, reading is influenced b!. the formation water.

Consequently, the preferred met hod of estimating movable oil in fresh rriud is with the electromagnetic propagation technique describrd in chapter 5 .

SUMMARY

ARCHIE RELATIONS FOR WATER SATURATION

General: S, -= c . +,IR, / 6

R, = deep resistivity, ohm-m R,, = interstitial water resistivity, ohm-m 4 = porosity, fraction c = 1 .O for carbonates, 0.9 for sands

Specific: S, = JROEJ

R, = resistivity of water-bearing formation of same porosity and R, as for R, forrnation

32 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION

RECOVERABLE OIL , STOCK-TANK BBL

(c 1 N = 7758.r. 4 e (1 - S,) . h . A/B r = recovery factor, = 0.2 h = thickness of producing formation, ft A = well spacing, acres 5 =;- volume factor of oil, z 4 2

RECOVERABLE G A S , MMscf

r = recovery factor, = 0.7 P. = reservoir pressure, psi T, = reservoir temperature, O F

Z = gas deviation factor, 0.8-1.2

G = 1.54 . r * <b . (1 - S,) . h . A . P,/[ 2 . (460 + T I ] (dj

Track 1 (SP and/or GR); pickout perme-

ty cwve in Track 2 (usually Induc- tion) in permeabiezones, looking for high values. These musf be due either to hydrocarbons in pores or to low porosity.

3. Read porosities in zones of interest (Track 3 or separate log). 4. If there is a nearby water sand of the same porosity, apply

Eq. b to obtain water saturation. 5. If step 4 does not apply, determine R, b y applying Eq. ato

the nearby watersand with S, = 1. Then apply Eq. aagain to the zone of interest to obtain water saturation

6. Calculate recoverable hydrocarbon b y E q s cor á.

VASIO IO^ Can correct R,,,, to R, with three resistivity curves.

* Drilling mud may be fresh (Rmf > I?,) or saline (Rmf < R,]. * Movable oil can be calculated

S,, - S, = c (ti~,,i~,, -./Rwin,) 156 (e)

R,, = mud filtrate resistivity at level of interest, ohm-m


R,, = flushed zone resistivity, ohm-m.

The calculation is more reliable with salt mud than fresh mud.

REFERENCES J t i V -

ity of Brine-Saturated Sands in Relation to Pore Geometry,” AAPG Bull., Vol. 36 (February 1952), pp. 253-277.

“.E. Archie, “The Electrical Resistivity Log as an Aid in Determining Sorne Reservoir Characteristics,” S P E - A M E Transactions, Vol. 146 (1942), pp. 54-62.

3M. Condouin and A. Heim, “Experinientally Determined Resistivity Profiles in Invaded Water and Oil Sands for Linear Flows,” SPE paper 712 presented October 6-9, 1963.

‘C Boleldieu and A. Sibbit, “A More Accurate Water Saturation Evaluation in The Invaded Zone,’” SPWLA Logging Symposium Transactions (June 1981).

’ Chapter 3

PERMEABLE ZONE LOGS he first step in analyzing a set of logs. as outlined previously. i s to pick out the permeable zones, lvliich may be sands or carbonates. and discard the impermeable shales. The logs used for this purpose are

the Spontaneous Potential (SP) and tfic Cainnia Ray (GR). Thcy are always recorded in Track 1.

The two logs distinguish shales from nonshales by quite different niecha- nisms. The SP is an electrical measurement and the GR is a nuclear iiieasure- inent. Sometimes the logs are virtually identical: sometimes the!, are vastly different. Fortunately, when one is poor, thc: other is usually good.

Fig. 3-1 compares SI’ and GR logs in a t!.pical soft rock sand-and-shale seqiience. Both curves are good in tliis case and clearly distinguish the shales on the right from thc perrneable sands on the left. Iri soft rock tlie SP gmerally gives a more black-and-white distinction bet\veen tlie shalps and the sands than does thc G H . The latter shows more variability in both shale and sand readings.

By contrast, in hard limestone formations the SP niay be a lazy, poorly developed curve that hardly resolves permeable and impermeable zones. The Gil is siiperior under these conditions, giving good shale-carlmnate distinction and bed resolution.

Both curves are tised to indicate the shale content of a permeable zone for shaly formation interpretation (chapter 7 ) . The GI1 i s inore quantitative than the SP in this respect. On the other harid the SP may be used to give the formation water resisti\ity reqiiirrd for saturation calculations.

~

I

_ _ _ _ _ _ ~ - -

SPONTANEOUS POTENTIAL (SP) L O G

The SP log is a recording of the difference in electrical potential bctwee~i a fixed electrode at the surface and a movable electrode in the borehole. The liok must be filled with conductive mud. N o SP can be measured in oil-base mud, ernpty holes, or cased holes. The scale of the SI’ log i s in millivolts. There i s no absolute zero; only cliarige~ in potential are recorded.

35

40 ESSENTtALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION

ior is observed. At very shallow depths where formation water is fresh, the SP in sands is reversed, i.e., has positive polarity relative to the shale value. Slightly deeper, perhaps at about 1,000 ft, the SP goes to zero. As depth progresses the formation water gradually becomes more saline and the SP increases in magnitude (negatively). Values of 70-100 niv at depths of 8,000 ft aFM 41 greatw ~ e ~ ~ s the eo te water sornci'rimcs es again in salinity, especially when overpressured formation5 are encountered. In such cases the SP correspondingly reduces in magnitude. In unu- sual cases it may even reverse again, as Fig. 3-3 illustrates. The SP in the permeable bed at 6,300 ft is normal; the SP at 9,100 f t is reversed.

In typical salt muds the SP is often useless because the SP magnitudes at depths of interest are small (Rft,* = IR,) and because boundary definition with low-resistivity mud and high-resistivity formations is extremely poor. This is explained in the following.

SHAPE OF THE SP CURVE

The measured SP is really the potential change that occurs in the borehole as a result of SP-generated currents flowing through the resistive borehole fluid. This is illustrated in Fig. 3-4. Opposite a shale far from a permeable zone boundary, no current flows, so the potential is constant. As the boundary of a pe e Zone is approached encountered, which causes the pot go negative with res shale.

Directly opposite the boundary, current flow is maximum so the potential change per unit length of hole, which is the slope of the SP curve, i s greatest. Beyond the boundary the current density decreases and gradually goes to zero. If the bed is thick enough the potential becomes constant very close to the SSP value. This value is always measured relative to the shale base line. As the other boundary with the shale is approached, the reverse situation is encountered and the potential returns gradually to the shale value.

The sharpness of the SF curve at a boundary, and therefore the vertical bed resolution, depends on the pattern of current flow at the boundary. This is a complicated function of the relative resistivities of the mud, permeable zone (invaded and noninvaded portions), and shale and to a lesser extent is a function of hole diameter and depth of invasion. The overridingprinciple is that currents seek the lowest resistance path in a loop. In a given formation they will spread out to the point that the formation resistance, which decreases as I/area of flow, is negligible with respect to the borehole resistance, which increases linearly with length of current path in the mud.

PERMEABLE Z O N E LOGS 44

Consequently, when the ratio of formation to mud resistivity is high, the currents will spread widely. There will be long flow paths in the borehole, and bed boundaries will be poorly defined. Conversely, when that ratio is

¡ I RESiSTiY b H M 5 k ITY M

I - I

I I

, 2 0

-A-

Fig. 3-3 Example of normal and reversed SP deflections


U j U

PERMEABLE Z O N E L O G S 43

h v , 1)ounclaries will he sharply defined. Note that the boundaries are opposite the points of inflection (maximum slope) of the SP curve, not necessarily opposite the Ii alfway point between shale and permeable-zone SP levels.

The more gradual the boundary transition on the SP, the poorer the bed resolution \vil1 be. A rule of thunib is that the bed resolution 11 (ft). defined as that thickness requirecl for the SI' to reach 80 ";o of the thick-bed reading (tlie SSP), i s given by

(3.1)

\vliere R, is tlie resistivity of the iiivadcd zone (the shallow log reading) and R,, is the mud resis.ti\ it!. ,4n even coarser approximation is

11 = li4J (3 .8)

Tvhere 4 is the porosity (fractional). This indicates that pernieable bed resoliition varies from a h i t 3 ft at 30%3 porosity to 30 ft at 3 % porosity.

SP logs therefore d~l ineate permeable beds quite well in porous cand- and-shale sequences but resolve beds very poorly in tight formations. In hard rock areas there may be massive low-porosity limestone or ~nhydrites of x.ery high resistivity. These show l ip as long. straight lines (no current at all entering or leaving the hole opposite these formations) on tlie SP and make interbedded permeable zones rt:cognizable only by clianges in the slope of the lazy SP cur\'e. Fig. 3-5 is an illustration.

Senice company charts are available to correct the SP response in a thin I ) e ~ l , ~ . ' " such as zonc B of Fig. 3-6, to \chat it woiild b e in a thick twi i f there arc no thick l>eds to indicate the SSP directly. IIo\vever, thcse charts are riot too precise. A correction factor greater than 1 .3 has diibioiis accuracy.

COMPUTATION OF R, FROM THE SP

Eq. 3.6 is used exteiisi\dy to determine the formation water resistivity that is required for \vater saturation calculations. First. the SSP is read from the log as the difference in millivolts t)et\veen the shale level and the thick- clean-water-said level nrar the zone of intercst. This i s illiistrated in Fig. -3-6, The shale line is taken as the maximum SP excursion to the right. The sand line is taken as the maximum deflection to the left in zone A , which is a \\ ater-bexaring sand. as shoxvri \)y the l o ~ v reading (1 ohrii-m) of the deep resistivity (dashcd) ci1n.e. The SSI' is read as - ti8 ni\', thtl scale tjeiiig 10 niv pt'r chart di\.iaion as indicated i n the log heading.


In such cases the dashed lines in the lo\\ er half of Fig. 3-8 should be used to transform R,,, to R,, , These lines are for average fresh forniation waters; corrections can be large. For example, R,,,,, = 0.7 at 100°F leads to R, = 1.4, a factor of two increase.

Whenever possible the R,, value obtained from the SP should be checked Eq. 2 ,, -= 1 i n therante or nearby

water sands. If there is good agreement, \vater saturation in hydrocarbon zones can be calculated with confidence. If there is substantial disagree- nient, the value derived from Eq. 2.8 i \ preferred. However, the reason for

001 5 0 0 ° F E K--400°F

Fig. 3-8 Determination of R, from R,, (courtesy Schlumberger)

PERMEABLE Z O N E LOGS 49

the discrepancr should be sought, particularly by checking the R,,, nieasure- ment and if possible obtaining a sample of formation water for analysis.

SP LOG IN SHALY SANDS'2s'3

\VI1

clay particles create internal membrane potentials that, \vhen added together, constitute a potential opposing the normal electrochemical potential in the adjacent shale. This reduces the SSP to a pseudo- static value called PSP. C'rider ideal conditions where the shale laniinations have thr same resistivity as the \arid laminations (both invaded and iinivaded portions), the percent reduction in SSP equals the percentage of shale by volume.

In the case where the sand is suttstantially more resistive than the shale, the percent reduction in SSP i s much greater than the percentage of shale.

oil-beating zone by virtue of it9 higher resistivity. The SP log is used as one of the prirne indicators of shale fraction, V,,, in a

shaly sand. Computation of Vq!* i s descrihed in chapter 7. The value obtained from the SP tends to be an iipper limit for the reasons just indicated.

SP ANQMALIES DUE TO VERTICAL MIGRATION O F FlLTRATE '' !!'lien a her)' permeable sand containing salt \vater is invaded hy fresh-

rtiud filtrate, the lighter mud filtrate will float up toward the upper bound- ar>'of the sand. After a few dn)s, inva\ion \vil1 be quite deep just below the tipper boiiridar!. and ver) shallow near the louer. The SSP will be signifi- cant11 reduced at the bottom because the diffusion potential, E,, disappears (n ith no invasion). Further, it is replaced b) an electrochemical potential acrojs the mud cake that now directly separates the borehole miid and the formation water that has reclaimed the originally invaded pores. This potential opposes the normal shale potential Es,,. Both effects reduce the normal SP deflection at the bottom of thesand. At the top of thesand, the SP deflection will be rounded due to the higher resistivity of the deeply invaded fresh filtrate.

50 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION PERMEABLE ZONE L O G S 51

Further, if such a sand bed contains thin, interhedded shale strcak5, the SP will show negative and positive undulations from the normal level above and below each streak.

SP ANOMALIES DUE TO NOISE Spurious DC currents in the ground and in the boreliolccan be causcd by

telluric currents, i)imetallic potentials, cathodic protection del ices, leaky rig power sources, welding machines. etc. Other sources of noise are magne- tizcrl cable drums and intermittent contact5 between casing and logging cable. Occasionally these niay disturb the normal SP, biit iiTiially they caii be eliminated with appropriate nieasiires.

THE GAMMA RAY (GR) LOG

The basic CR log is simple arid easy to record and to understand. It was routinely recorded for many years with only minor improvements in instrumentation and with limited quantitative use. In the past few years, ho\x.- ever, the introduction of the spectral CR has given the recording of earth radioactivity a new lease on life. It is exciting to see new uses for the technique unfold.

BASIC GR LOGS The basic GR log is a recording of the natural radioactivity of forma-

tions. The radioactivity arises from uranium (U), thorium (Th), and potassium (I<) present in the rock. These three elenitmts continuously emit gamma rays, which are short bursts of high-energy radiation similar to X- rays. The gamma rays are capable of penetrating a few inches of rock. A fraction of those that originate close to the borehole traverse the hole and can be detected by a suitable gamma-ray sensor. Typically, this is a scintilla- tion detector, 8-12 in. in active length. The detector gives a discrete electrical pulse for each gamma ray detected. The parameter logged is the riurriber of pulses recorded per unit of time by the detector.

GR logs are scaled in API units (APIU). An APIU is 11200 of the response generated by a calibration standard, which is an artificial formation containing precisely known quantities of uranium, thorium, and potassium maintained by the American Petroleum Institute (API) in Houston. l í 'The response generated by this formation is defined as 200 APILT. By design the

h\c95 3 *>vLrLc5 => L., $ T h y k Mi5 .J?&C'j ih12 L~ : it & +h,5 d. ,ohe' l \ L*b~kt,a\ G t '>~avhx.\aq

calibration standard has t\vice the activity of an averageshale, considered to contain 6 ppm (parts per million) uranium, 12 pprn thorium, and 2% potassium. Consequently. shales read in the vicinity of 100 APIU on CR logs.

Response to Different Formations

Gamma Ray logs are effective in distinguishing permeable zones by virtueof the fact that the radioactive elements tend to beconcentrated in the shales, which are impermeable, and are much less concentrated in carbonates and sands, which are generally permeable. Fig. 3-9 sliows typical responses. Limestones and anhydrites have the lowest reading, 15-20 APIU; dolomites and clean sands have slightly higher values, about 20-30 AI'IU.

Shales average about 100 AI'IlJ but can vary from 75 to 150. A few very radioactive shales--the Woodford, for example-niay read 200-300 APIU. Normally, therefore, the GR log separates clean sands and carbonates from shale quite nicely. However, there are localized areas where sands and dolomites, even though fairly free of clay, are radioactive enough that distinguishing them from shales is difficult. Among the less commonly encountered forniations, coal, salt, and gypsuni give quite low readings; volcanic ash and potash beds give high readings.

Depth of Penetration and Vertical Resolution

The depth of penetration of the GR log is 6-12 in.. being somewhat higher at low formation density (high porosity) than at high density. Verti- cal bed resolution is about 3 ft. It is dependent on logging speed. as explained later.

Borehole Effects

The GR log is calibrated under conditions of &in. hole, 10-lb niud, with the logging tool (3Ys-in. diameter) excentered in the hole. Nocorrections are required for these conditions. With larger hole sizcs and heavier muds or with the logging tool centralized, there is more gamma-ray-absorbing material between the formation and tool, and the response drops. Con- versely, the response is increased in smaller or empty holes. Correction curves may be found in service company chart books.

Correction factors are normally modest, in the range of 1.0-1.3. 'They can be ignored in hand interpretation except in the combination of circumstances where the GR is being used for determination of shale content, the shales are badly caved relative to the sands. and the mud weight is very high. The corrections are also important in the infrequent cases where the GR log is used to assay potash or uranium deposits.

52 ESSENTIALS OF MODERN OPEN-HOLE LOG INTERPRETATION PERMABLE ZONE L O G S 53

O 5@ 100 API units I 1 L

Shaly sand

Shale

sand

f Clean limestone t Dolomite

Shaie

Fig. 3-9 Gamma Ray response in typical formations

Occasionally the mud in a well contains excessive amounts of potassium or uranium, either because potassium chloride has been added (to prevent shale swelling) or because potash or uranium beds have been penetrated. The potassium or uranium in the mud will contribute to the GR log, giving an abnormally high background level (that will vary with hole diameter) on wh mal k e S

method of correction for this effect has been published."

Shale Determination

Because uranium, thorium, and potassium are largely concentrated in clay minerals, the CR log is used extensively in shaly sand interpretation to estimate the fraction of shale by volume, V,,,, in thesand."This procedureis described in chapter 7. Basically, it is a matter of estimating the clean sand and 100 70 shale levels on the log and interpolating between them to determine V,, in a partially shaly interval. It is not a very precise technique, so other shale indicators are used as well.

SPECTRAL GR LOGS'8s'9

The radioactive elements uranium, thorium, and potassium emit gamma rays of different energies, as shown in Fig. 3-10. Potassium single energy at 1.46 rnev (million electron volts). Thorium and ura emit gamma rays of various energies, the major distinction being a pronii- nent thorium energy at 2.62 mev and a predominant uranium energy at about 0.6 niev. In principle it is therefore possible to distinguish the three emitters by anal? zing the energies of detected gamma rays.

As gamma rays progres\ from the point of origin in the formation to the detector in the borehole, their energies suffer severe degradation. Further smearing takes place in the detector itself. Nevertheless, it is possible with appropriate instrumentation and careful analysis of the spectrum of pulse amplitudes from the detector to break down the total GR log into its uranium, thorium, and potassium components and to generate spectral GR logs showing directly the concentrations of each of these elements in the logged formations.

Fig. 3-11 is an example. Uranium and thorium are scaled in parts per million (ppm) and potassium is scaled in percent by weight jwt%). The three elements contribute roughly equally to the total counting rate even though potassium is present in much greater concentration. Thespectral log shows large differences in the shales above and below the clean zone from

I

1

4

L I

I

[o

K

a, 2.

L

O

in O

u,

O

in N

N

O

u,

O

O

O

?

.-


eliminated are better indicators of shale content than the total GR activity. This is because uranium salts are soluble and can be transported by liquid movement after primary deposition. They may be picked up in one location and precipitated elsewhere, particularly where a pressure drop occurs. Fig. 3-12 shows a case where the standard GR log indicates interval A to be three St? ed b the thorium and potassium spectral curves clearly show the reservoir to be a continuous unit. The shale streaks are simply permeable zones high in uranium content. New wells in old fields often show such response opposite permeable zones that are producing in adjacent wells.

In other areas where abnormal amounts of potassium-bearing micas and feldspars are associated with the sands (the North Sea, granite wash areas of the southwestern U.S.), the thorium component alone is the best indicator of shale content. In any case, whichever shale indicator is used, thorium- potassium or thorium alone, the shale fraction is derived from the desired spectral curve in exactly the same fashion as described in chapter 7 for the total GR curve.

Efforts are underway to determine from spectral GR logs the type as well as the amount of clay encountered in formations. Fig. 3-13 is a proposed

I I lo API UNITS 1501 I K O o 5% co

Th 0 7.0 ppmlCD

1 1 I

A

Fig. 3-12 Spectral GR log showing single reservoir instead of three iT: interval A (Alberta) (couriesy Dresser)


I 917 / /

POSSIBLE 1000 KAOLINITE . l o o m ILLITE ’POINT’ MONTMORILLONITE,

20

1

i

O 1 2 3 4 5 6 K(%)

Fig 3-13 1denttficat:on of clays and other minerais from potassium and thorium responses (courtesy Schlumberger, Q SPE)

chart for identification of clay type from thorium and potassiuni log resporises.”Approaches such as this will assist in determing formation productivity, since different types of clay have quite different surface-to- volume ratios and therefore .water-retention and permeability-reduction capabilities.

In another application the uranium component of thespectral curves has been used to indicate fractured zones in tight carbonate?.’.‘ High uranium content is taken ab an indication that fissures in the zones were permeable enough to flow fluids in the past and therefore may still be able to conduct oil arid P ~ S from rcrriote porosit! .

STATISTICAL F L U C T U A T I O N S Ganima Ray logs never repeat exactly. The same is true for all types of

nuclear logs. Some of thesmall wiggles on the logs are statistical fluctuations that do not reproduce and do not represent formation variations. In reading the logs, averages over about 3-4 ft should be used. The exception is a bed that is less than 3 ft thick, in which case the peak reading should be taken.

The Source of statistical fluctuation is the random nature of nuclear processes. The emission of any one gamma ray in a formation is in no way time related to any other. Consequently, the pulses from the gamma ray detector appear as a random sequence. The number occurring in any given


time interiaal \vil1 differ from that in a succewive but identic.al tinw interval, ei’en thoiigh the detector is stationary. I’ercciitage-\vise this difference will be small if the timc interval is long enough that a large nilinher of p u l w s is ci«untect. Conversely, it nil1 be large if the averaging t ime is small. A measure of the percentage fiuctiiation is 1 0 0 i 6 where N is the atwage niiniber of piilses in the measiiring interval.

Typical pulse rates for the basic CH log are 50-300 pulses per second. The iisual averaging time is 2 sec. This means the log will show fluctiiations of about 4 (jó above and helow the mean reading in shales (N = 600) and 1 0 % either side of the mean in clean sands or carbonates (N= 100). ribsoliite magnitudes will be in the range of k 5-10 API units in shales and k 2-4 API units in clean forniatioris. The best way i o appreciate this is to overlay repeat logs.

Statistical fliictuations may he reduced by increasing the awraging time, but the valrie employed sets a limit on the logging p e d . Fig. 3-14 shows the approximate response (without statistical fluctuation) to a 4 ft bed ui th a detector of 8-12 in. active length for a 2-sec averaging time and for logging speeds of 600, 1,800, and 5,400 ftihr. At 600 ftihr the boundary response is primarily determined by the length of the counter; resolution will not improve at slower speeds. At 1,800 ftihr the bed is adeqiiately resolved, althoiigh it appears to be shifted about G in. in the direction of logging; compensation is made in the recording process for this shift. At 5,400 ftihr the bed is quite distorted and its amplitude is reduced us well as being shifted about 1s in.

If the averaging time is doublcd, for example. logging speed must be halved to niaintain the same bed resolution. A Compromise must al\\,aq.s be made bctween log acciiracy and logging speed. A reasonable logging speed is that whirh nioves the detector its effective length (1 ft) in the averaging interval. For 2-sec averaging this is 1:800 ftihr: most nuclear logs are rim at this speed. If the speed varies, bed resolution will change but the iriaqriitutles of statistical fluctuations will rcniain constant.

With logging-truck computers, averaging over a fixed depth interval rather than time interval may be employed. Typically this is set at I ft for the GI1 log. With loggingspeed of 1,800 ftihr, effective averaging time is again 2 sec. However, if logging speed is doubled, averaging time is reduced by a factor of 2. Bed resolution remains fixed but statistical variations increase b y a factor of v’, .

The problem of statistical uncertainty is aggravated mith spectral CR logs because the counting rates in the uranirini. tlioriiini. and potassium channels are 3-10 tinies lo\ver than that of the total GR, depending on the


APlU

O

jp \

\t‘5’400 f‘ hr

\

120

- Fig 3-14 Effects of G R weraging t i r n 6 1 and logging weed on bed resolulion

60 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION PERMEABLE Z O N E L O G S

particular instrumentation. This means averaging times must be increased and logging speed must be slowed. Typical values are 4-6 sec and 900-600 ftihr. These are painfully slow logging speeds, so the best technique i s to first log the well at normal speed while recording a total GR curve and then rerun the short sections of vital interest recording the spectral curves.

SUMMARY

SPONTANEOUS POTENTIAL L O G 0 A recording of voltage generated b y electrochemical and

electrokinetic action at the junction of a permeable zone and shale.

0 Distinguishes impermeable shales from permeable sands or carbonates.

op with only a fraction of a miilidarcy, but there is between magnitudeand absolute permeability

0 Shales appear as excursions to the right, permeable forma-

s are at points of inflection, not halfway points, good at high porosity, poor at low porosity.

0 Magnitude depends on contrast between R,, and R,. Thus, SP delineates permeable zones well in fresh mud where R , , >=. R, but poorly in salt mud where R,, = R,,.

0 I?, obtainable from SP, but values must be chosen carefully, especially at depths at less than 3,000 ft.

0 Shaly sands reduce SP especially when hydrocarbon bearing Maximum value of shale content can be computed from iog

* Abnormal readings obtained in pressure-depleted zones and in high vertical-permeability zones.

GAMMA R A Y L O G 0 A measurement of gamma-ray intensity in the borehole due to

U, Th, and K tend to concentrate in shales and occur least in natural disintegration of U, Th, and K

clean sands and carbonates.

Shales appear as excursions to right; clean formation to the

GR curve is good in hard rock regions where SP is deficient. A prime indicator of degree of shaliness of formation. Shaliness estimated b y interpolating between clean formation line and

Spectral GR logs separate total response into U, Th, and K contributions. Improves reservoir delineation and shale estimation,

0 Do not repeat exactly. Statistical fluctuations limit speed to 1,800 ftihr (vs 5,000 ftihr for Sonic or Electric). Also holds true for all nuclear logs. Do not read sharp peaks. Averages over 3-4ft should be taken. Repeat run overlays indicate magnitude of statistical fluctuations.

I left.

I 1

Depth of penetration-6 in.; vertical resolution - 3 ft.

REFERENCES H. G. Doll, “The SP Log: Theoretical Analysis and Principles of Interpreta-

‘ M . R . J . Wyllie, “A Quantitative Analysis of the Electro-Chemical Compo- h. , Vol. 1 (1949), p. 17. id h4.H. Waxman, “An Investigation of the

B1.R.J. Wyllie, “An Investigation of the Electrokinetic Component of the

hl Gondouin and C. Scala, “Streaming Potential and the SP Log,” Jour

(’ M.R.J. Wyllie, A . J . deWitte, and J .E. iriarren, “On the Streaming Poten-

Ii. J. Hili and A.E. Anderson, “Streaming Potential Phenomena in SP Log

Voy E. Althaus, “Electrokinetic Potentials in South Louisiana Tertiary

tion,” T?-ansAIME, Vol. 179 (1948).

nerit of

Electro-Chemical Component of the SP Curve,” Jour. Pet. Tech. (March 1962).

Self-Potential Curve,” Trans. A M E , Vol. 192 (1951), p.1.

3

P P t . TP4.h. (August 1958).

tial Problem,” Traris. .41ME, Vol 213 (1958), p. 409.

Interpretation,” Trans. AIME, 1‘01. 216 (1959), p.203.

Sediments,” The Log Analyst (May-July 1967), p.29. ‘ Schfiimberger, Log Interpretation Charts (1979). lo Dresser Atlas, Log Interpretation Charts (1979). I’ M . Gondouin, M.P. Tixier, and G.L. Simard, “An Experimental Study on

the Influence of the Chemical Composition of Electrolytes on the SP Curve,” ]our. Pet. Tech. (February 1957).

H.G. Doll, “The SP Log in Shaly Sands,” Trans. A M E , Vol. 189 (1950). l3 L. J.M. Smits, “SP Log Interpretation in Shaly Sands,” Soc. of Pet. Eng. I . ,


l 4 F. Segesman arid M.P. Tixier, “Some Effects of Invasion on the SP Cunr:.” J o u r . Yet. Tech. (June 1959).

l 5 M’.B. Belknap, J.T. Dewan, C.V. Kirkpatrick. \?’.E. hlott, A.J. Pearson, and W.R. Rabson, “API Calibratiori Facility For Nuclear Logs.” Zlriíí. atid Prod. Prac. (Houston: API, 1959).

l6 J,LV. Cox and L.L. Raymer, “The Effoct of Potassium-Salt Muds on Carnrna Ray and Spontaneous Potential hleasurenieiits.“ SZ’WLA Logg ing S!y7t2po- .siurti Transactions (June 1976).

A. Heslop, “Gamma Ray Log Response of Shaly Sandstones.” Thc Log iinalyst (September-October 1974).

C.A. Lock and W.A. Hoyer, “Natural Gamma-Ray Spectral Logging,” Tlie Log Analyst (September-October 1971).

Dresser Atlas, “Spectralog” (1080). M . Hassan, A. IIossin, arid A. Combaz. “Fl.indanientals of the Differmtial

Gamma Ray Log-Interpretation Technique,” SPJC’LA Logging Syrnposiiini Trans- artion.? (June 1976).

O. Serra. J.L. Baldwin, arid J . A . Quircin, “Theory and Practical Applica- tions of Natural Gamma Ray Spectroscopy,” SPU’LA Logging Symposiiln~ Tronsac- tiotis (July 1980).

22JV.H. Fertl, and E. Frost, Jr., “Experiences uith Natural Camnia Ray Spectral Logging in North America,” SPE 11145. New Orleans (September 19S2).

J.A. Quirein, J.S. Gardner. and J.T. Watson, “Combined Natural Gamma Ray SpectraliLitho-Density Measurements Applied to Complex Lithologies.” SPE 11143, New Orleans (September 1982).

17

18

23

24 Dresser Atlas, “Spectralog,” ibid.

Chapter 4

RESISTIVITY LOGS he basic interpretation relation in well logging, as dcvelopetl in chapter 2, is the water saturation relation T

The most important input to this equation (since its value can never be guessed) is the resistivity, R,, of the iiiiinvaded region of the formation in qiiestion.

No resistivity-measuring tool has yet been designed that can reach deep enough to guarantee reading 11, under all possible invasion conditions while retaining good bed resolution. Therefore, from early days resistivity logs have consisted of three curves: deep, medium, and shallow investigation. With these three measurements and the assumption of a step invasion profile, correction can be made to the deep reading to obtain R,.

Nevertheless, many logs have been run with only two curves, deep and shallow reading. These clearly show invasion effects but do not permit a correction to the deep reading, which must be assumed equal to 11,. The assumption is reasonable in high-porosity areas where invasion is shallow hut can lead to significant errors in low-porosity regions where invasion may be deep.

Over the years there has been a continual succession of resistivity tools with improved designs replacing older ones. It would \)e convenient to forget the obsolete versions, but we cannot. Company files and lug li!xarics still abound with old logs that are continually being reviewed for new drilling or production prospects.

CLASSIFICATION AND APPLICATION

Table 4-1 is a classification of the major resistivity tools that h a w been rivxi or are in use. Tlirl curves are categorized b y their radii of investigation, i.e , deep(3+ f t ) , mediiirn (1.5-3 f t ) , Challow(0.5-1.5ft), anclflushcdzone

63

Sal

t m

ud

Rrni or

R, *

200

Flus

hed

Zone

1

-4 in

.

Mic

rolo

(M

L)

Min

jog

- C

onta

ct)

Pro

xim

ity (

PL)

Mic

rola

tero

log

(MLL

! (F

oRxo

)

‘MLL

or

FO

RG

-

Mic

ro

Sph

eric

ally

,F

ocus

ed

A

CLA

SSIF

ICA

TIO

N O

F RE

SIST

IVIT

Y TO

OLS

1965

-

1975

-

1972

-

Des

igna

!ions

C

omm

ents

ES, E

L 1 obs

olet

e

obso

lete

IE

S, I

EL

ISF

phas

ing

out

DIL

-LL8

, D

IFL,

DlS

G

1 D

IL-S

FL o

r cu

rren

t D

lSF

~

LL-7

1 o

bsol

ete

LL-3

gua

rd

<.SLb$

kk, j cu

rren

t

I

DLL

-MS

FL

curr

ent

I

I I

le

Num

bers

are

50%

radi

us o

f in

vest

laat

ion

mea

sure

d fro

m b

oreh

ole

wal

l

h

-Q

&a ESSENTIALS OF M O D E R N OPEN-HOLE LOG INTERPRETATION

with the Spherically Focused shallow log, the DISF, was introduced by Schlumberger in the mid 1970s. Gradually the Dual Induction tools are replacing the single Induction systems because of their invasion-correcting capabilities.

indicates the presence or absence of mud cake in great vertical detail. Therefore, it is an excellent permeable-zone or sand-count indicator-the best available. Under favorable circumstances it can give the flushed zone resistivity, R,,, but is not really designed for that purpose. The Microlog is discuqsed in chapter 5 along with the Electromagnetic Propagation Log, with which it has been recently combined,

The Proximity log is a focused curve that measures flushed zone resistivity, R,,,, in the presence of thick mud cakes that can occur with fresh mud. The R,,value can be used for m

nformation about the in rately from the Induction logs but can be run simultaneously with the Microlog. It has been superseded by the MicroSFL, which can be run simultaneously with the Induction,

SALT MUD TOOLS

Two medium-investigation focused tools were introduced in the 1950's for salt mud surveying. They were the Laterolog-7 and the Laterolog-3, also called the Guard log.' At the same time a flushed zone tool called the Microlaterolog or FoRxo, was introdiiced.' It could provide good R,, values for mud cakes up to 34 in. thick. The medium and flushed-zone curves had to be run separately and could only provide good water saturation and movable oil values if invasion was not too deep.

As a corisequence the separate tools were succeeded in the 19705 by Dual Laterolog systems comprising deep and shallow curves run simultaneously with a flushed zone log.'" Designations for Dresser and Welex are DLL- MLL and Dual Guard-FoRxo. In the case of Cchlumberger, the Dual Later- olog is combined with a Microspherically-Focused curve that can read R,, accurately over a wider range of conditions than can the MLLlFoRxo curve. The combination is denoted DLL-MSFI,. Once again the single Laterolog tools are being phased out in favor of the dual systems with their invasion- correcting capability.

REStSTlVlTY LOGS 69

RANGES OF APPLICATION OF INDUCTION LOGS AND LATEROLOGS

The Induction combination is the onesuitable in the majority of cases. It

whenever R,T,,/R, is greater than approximately 2 and the formation resistivity, R,, does not exceed about 200 ohm-m; the log is not accurate at higher resistivities. Resistivity increases as porosity decreases, so there is a low-

30

25

20

I5

IC

5

Fig. 4-3 Ranges of appiica Schlum berger)

70 ESSENTIALS O F M O D E R N OPEFI-HOLE L O G INTERPRETATION 1 RESISTIVITY L O G S 71

I porosity limit. The cutoff corresponding to 200 ohm-m is dependent on the R, value, as indicated. Typically, R,. is about 0.05 and R,,,,/R, is approximately 10 in fresh mud, in which case the low-porosity cutoff is 5 % .

If the mud is conductive relative to forniatiori \vater, i.e.. R,,,/R,, < 2, tlie Laterolog should be run. It is also the best in fresh mud when resistivities above 200 olinis arcs encountered because it is accuratc at high resisti\.iiies. In borderline cases, large boreholes (> 12 in.) and decp invasion (I> 40 in.) favor the Laterolog. This cominon1)- occurs in the Eastern Hemisphere.

The reasons why the Induction and Laterolog are preferred under the indicated conditions will become apparent in the next two sections lvhere the principles and interpretation of the tools are presented. The discussion is confined to the Schlumberger conihiriations. the DI ISFL on the one hand and the DLL-MSFL on the other, since they are the most advanced tools in each category. If these are iinderstood, there is no difficult!, i n interpreting logs riiri with older or single-curve versions of Induction or Laterolog tools.

DUAL INDUCTION-SPHERICALLY FOCUSED LOGS

The Induction log inherently wnses the conductility of the formation, which i s the inverse of its resistivity. In conimonly tised units

conductivity in rnmho/rn = I ,000iresistivity. ohm-ni

‘The principle of the Induction system is illustratcd i n Fig. 4-4.”.’ A constant current of 20 kHz frequency is fctl to a transmitter coil. This generates an alternating magnetic field that causes a circular current ( F o u - cault or eddy current) to flow in the surrounding medium. This current in turn creates a magnetic field that induces a voltage in the receiver coil. The induced voltage is approximately proportional to tlie surroiiriding conductivity. From this voltage the formation conductivity and thence its resistiv- i!y is derived for presentation 011 the log.

With the single-transmitter, single-receiver system shown, coiitrihu- tions froni the borehole and invaded zone as well as from neighboring beds above and below the coil pair would constitute a significant portion of the receiver signal. Practical Inductinri tools therefore iitilize an array of auxiliary transmitter and rcceiver coils, spaced above and below the niaiii ones.

to niinirnize tht.se contributions and to riiaxiiiiíze depth of penetration and vertical resolution. Typically, six or more coils with approsimatcly 40-in. spacing bet\veen tlie main tritnsniittt.r-receiver pair are used (Fig. ‘$.-5) to obtain the d r e p d reading curve, denoted IL,i. Fewer coils are used to provide the niediurn reading curve. denoted IT,,,,. The Induction log reqiiirts no condiictixve fluid in t h e borchole for its opera- tinn. It works very well--in fact, best- in holes filled with air or gas or with oil-liase niud. N o other resistivity tool can be used under thcse circuni- stances .

Depth of Penetration of Induction Logs Fig. 4-6 shows the depth of penetration of the deep, IL,. and mcdium.

IL,,,, arrays. The ititcgrateti gconietricaljnctor plotted is the relativi? weight that the tool assigns to a cylindrical shell of the surrouiiding mediiirii extcnding froni the surface of the sonde to any particiilar diarneter. Each shell contribiitcs to the total conductivity signal in accordance with the product of its conductivity and i t s relative weight. For example, if C,, is the

_ - I I.-

Fig. 4-4 Induction priricipie (courtesy Schlumberger)


conductivity of the invaded zone, d, its diameter, and C, the conductivity of the undisturbed formation, then the apparent conductivity, C,, read by the tool (neglecting any borehole signal) is

c, = C,&, $- Ct(1 - GdJ (4.2)

In terms of resistivity

(4.3) liR, = Gd,/Ryo + (1 - G,,)/R,

For homogenous formations M ith no invasion, Fig. 3-0 (solid line) gives relative signal contributions from different regions. For example, 50 70 of the signal comes from within and 50 % comes from beyond a diameter of 11 f t for the IL,. The corresponding figure for the IL,, is 5 ft, so i t has only half the depth of investigation as does the IL,. Note that the IL, assigns greatest

art of the forma that part a t 40

heir respective curves.

d at about 60 in. di r, as shown by the

n

- Fig. 4-5 Practical Induction array (courtesy Dresser]

RESISTIVITY L O G S 73

For the usual case of invaded formations, the above equations may be used to illustrate why the Induction works well in fresh mud and poorly in salt mud. Ega&y\~

Consider a typica fresh mud case where R,, = 1, R, = 10, R,, = 20, and d, = 65 in. The IL for 65 in. is 0.2. Eq. 4.3 gives

1/R, = 0.2/20 + (1 - 0.2)/10 R, = 11 ohm-m

The íL, therefore reads only 10 % in error. Applying a small correction factor obtainable from the lL,, and SFL readings is reasonable.

The situation with salt mud is quite different. The same formation, drilled with salt mud of R,, =0.05 ohm-m, would have R,,= 1.0. The IL,, reading would then be

1/R, = (4.211.0 + (1 - 0.2)/10 R, = 3.6 ohm-m

In this case the IL, reads far from the correct value. It would be impossi- ble to obtain a sufficiently accurate correction factor from the IL, and SFL curves.

I - INTEGRATED RADIAL GEOMETRICAL FACTOR .- u W u

a O i- U

LL a

10

O 8

0 6

04

-No S k i n E f f e c t

S k i n Effect i n c l uded : C a s e of 0 2

O O 40 80 120 160 200 240 280 320 3

I D I A M E T E R d i ( INCHES)

Fig. 4-6 Depth of penetration of Induction log (courtesy Schlumberger]


In reality the error in salt mud conditions is even greater than illiistratecl because the geometrical factor curves of Fig. 4-6 that apply to qiiite resistive formations shrink to srnaller diameters as resistivity decrcases. giving greater weight to the invaded zone (dashed line). This pheiiomenorr i s called skin or propagation cyffrct.

The important point is that the Iridiiction log will be adversely affected I)!. an invaded zone more conductive than the undisturbed formation. It prefers in\,aded zones less conductive (rnore resistivt,) than the iininvaded formation.

Borehole Effects

Borehole fluid contribiites a n iindcsircd sigiial to the Intliiction response. It is inconsequential when the n i i d is fresh. ihe hole size is small. or the formation conductivity is high. However, it can be significant iinder the re\trse conditions.

Borehole correction is made by rneari:; of Fig. 4-7. For the example illustrated, IL, with 1.5-in. standoff in a I4.6-in. hole, the borehole georriet- rica1 factor is 0.002. With R, = 0.35 ohm-m (2,857 mmhoim), the hole signal is (0.002 x 2,857) or 5.7 mmho/m, as indicated by the nomograph. This signal must be subtracted from the II,,, conductivity reading to obtain the corrected forniation conductivity. The correction is negligible if the II,, reading is less than 10 ohm-ni íi.e., greater than 100 nirnlioini) but significant at higher resistivities. For example, i f the IE,, read 50 ohm-in (20 mmhoim), thecorrected resistivit).woiild be 1,000/(20 - 5.7) or 7Oohrri-m.

The situation is aggravated Lvhen the miid is more conductive. For the same conditions u s cited, except with R , , = O. 15 ohm-ni. the hole signal is 13.3 mriihoini. An II,,, reading of 50 would correct to 149 ohm-rn, an excessively large adjustment. In general. a correction of niore than 50% to the indicated resistivity is unreliable because there is some uncertainty in the hole geometrical factor, the hole may be out-of-round. the standoff may h e affected by mud cakeor holerugosity, and thercmay besoine i~ncertaintyiri mud resistivity.

Fig. 4-7 shoLvs two additional important points. First, borehole corrections are much greatcr with the Induction tool against the wall than stood off from the wall. A minimum standoff of 1.5 in. should always be i w d . This is normal procedure for service companies but should be checked at the wellsite. Standoff devices are somrtirnes omitted whcri there is difficulty in getting the tool downhole.

RESISTIVITY L O G S

Second, the hole corrections are much larger for the merliiirn-reading IJ+ than for the JL0. In fact, corrections become escessivr Lvith hole diameters larger than 12 in., even \t.ith fresh mud . 'This limits the use of the dual Iiiduction tool in some areas.

Bed Thickness Effects The Induction log has a vertical bed resolution of approsimately 1 ft, as

determined primarily by the main coil spacing. There arc, however, coni ribiitions f r o m beds well above and helow the

4-ft section directly opposite the tool. These contributions arc negligible in soft rock but are significant in hard rock when the shoulders are miicli more conductive than the bed of interest. even when thc latter is iiiiich thicker

I - I 5

F l y 4-7 Borehole corrections for I L d and ILm (courtesy Schlumberger)


than 4 f t . Such could be the case for example of a 10-ft hydrocarbon-bearing low-porosity limestone sandwiched between two shales. The Induction could read too low by a factor of two in a worst case.

Charts exist in service-company manuals for bed thickness correction. However, they presume thick homogeneous shoulder beds, a condition not

thickness corrections are ignored, either as unnecessary or inapplicable to the conditions at hand.

Sonde Error

If the Induction sonde is lifted high in the air with no conductive niate- rial in its vicinity, there can still be a small signal on the order of a few mmhoim due to residual coupling between transmitter and receiver coils or to imbalance in the receiver circuits, This is termed sonde error. Normally the error is balanced out before the log is run. The log heading should so indicate. There is no guarantee however that sonde error remains the same

under high temperature and pressure condit ole effect, is the reason the

measuring formation conductivities less than 5 mnihoim (resistivities above 200 ohm-m).

THE SPHERICALLY FOCUSED L O G (SFL)

The SFL is the shallow-reading resistivity curve of the DIL-SFL combination. It is obtained with a set of separate electrodes mounted on the Induction sonde. Fig. 4-3 illustrates the principle. Survey current i,,, flows from the center electrode, A<,. A variable focusing current. i *, flows between '4, and the x c d i s r > . electrode pair, A, and Al ' , connected together. U t appropriate adjiistnient, the focusing current forces the survey current t o enter the formation in the same manner as it would if there were no borehole (with spherical equipotential lines such as B and C , hence the name). With this system, conflicting requirements of shallow formation penetration and independence of borehole size and salinity over a wide range are met.

The depth of penetration of the SFL is shown in Fig. 4-9. It i s significantly shallower than that of the predecessor curves, the LL-8 (or short Guard), and 16-in. Normal. This means it gives greater weight to the invaded zone, which is desired, but in general ítstill reads too deep to givean accurate measurement of flushed zone resistivity, Rxo.

The vertical bed resolution of the SFL, LL-8, and short Guard is about 1 ft. Bed thickness corrections are not required.


Borehole effects for the SFL are normally negligible. All shallow resistivity curves tend to read resistivities too low if the borehole becomes quite large and invaded zone resistivity becomes high relative to mud resistivity (meaning low porosity). For borehole diameters of 6-12 in., SFL corrections are negligible up to RsFL/R,, = 2,000. LL-8 or short Guard correction

cant if R,,/R,, > 30. Correctioti charts are found in service company chart books but are not often necessary.

bee CEt al imc; Signifia-

LOG PRESENTATION Fig. 9-10 shows a typical presentation of the 5-in.il00-ft DIL-SFL log

when it is run in combination with the Sonic log. The SP curve, which is obtained simultaneously, i s recorded in Track 1 on a linear scale. Also shown i s an R,, curve; this is described in chapter 9. The three resistivity curves are recorded in Track 2 and half of Track 3 on a logarithniic scale

W

Fig. 4-8 Principle of the spherically focused log [courtesy Schlumberger)

78 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION y

wvcring 0.2-200 ohm-ni. The deep curve is always heavy dashed, the nediuni curve is light dashed, and the shallow curve is solid: the latter is inaveraged (SFLU) in this case. In the center of Track 3 is the caliper ciirve dotted) with the bit size (7% in.) shown dashed. The Sonic log, to be liscussed later, is recorded in the right-hand side of Track 3. Note the breaks n the right and left edges of the grid. These occur at 1-min intervals and ,erve to indicate logging speed. In this c a e it W ~ I S 86 ftiriiin or 5,200 ftilir.

On the 2-in.1100-ft log the resistivities are presented on a linear scale to 'acilitate correlation with older linear scale logs. The SFL or I L 8 is aver- iged over 3 ft to reduce its detail to that of the Induction curves. Further, the jeep Induction conductivity reading is presented in Track 3 on a scale froni ight to left. This facilitates reading very low IL, values, such as may occur n salt-water sands. For example, ail IL, value of 0.55 ohm-m, difficiilt to -ead accurately o11 the resistivity scale, can be read easily as 1,820 mnihoim in the conductivity scale.

In Fig. 4-10, which is a log of the Travis Peak formation in East Tcxas, :he SP curve indicates permeable zones at A, C. E, and G and relatively

1 .o

O

d (inches) -+ -____

Fig. 4-9 Depths of penetration of the shallow resistivity curves (courtesy Schlumberaer)

RESiSTiVlTY L O G S 79

impermeable intervals a t 13. 13, arid i?. i n thc perrneal>le zones the three rcsisti\it), cur t~(~s shoxv wide separation indicating deep inva,;ioii. This i s supported by the caliper ciirvc urherc i i i i i d cakcs u p to ahout 2 i i i . i i i

ttiickncss are indicated (shaded sections). Zones L) and F slio\\. i w ciirve separation. indicating that they are foriiiationx. probably shales. ;\,itti zero pernieability. On the other hand interval n s1ion.s some separation 01' the S F L and Induction ciirvt"s hiit none between tiic two Induction logs. This

-ig 4-10 Typical presentcihon of Dual lnduciion StL (or LLS] log (courtes/ khiiJrnberger)


implies low but finite permeability with \hallow incasion. The section is a shaly sand or carbonate, not a pure shale.

I t is apparent from both the negative SP values and the R5FL > H l l n separation i n the permeable zones that the well was drilled with fresh mud, with H,,,, >> R,, . However, it is riot evident from the resistivity iog5 alone

e prc for reference. CZ” could calculate wa from the Sonic lug but will reseme this until later

Determination of R, The object of the interpietation procedure is to utilize the three resi\ti\ -

i t) values obtained m 2 given zurie to corrcct the deep rtJadiiig for in\ ds101-t

and thus to arrive at the best possibit‘ R. \ alue. This I \ dorie bl i i iedm of the tornado chart in Fig. 4-11.

The IL,, Ir,,,,, and SFL readings are first corrected for borehole anti bed thickness effects, if necessary, to give corrected H,,, R,,,, and R,, I 1 aluea

e location of this point, the \ err the dotted curves. (R,J

est.) For example, in the middle of zone C

ohm-m. Sacral poiiits are noteti orthy. First, the R, corrcction factor 15 I)eti\L-en

1 . O arid 0.6. althougli sonrcch;irts qo to O 3 This mean\ that (’\’en \\ ith drq) in\ a w n the II,,, reads closr to R,. I tool:, \\ ert s u c c w t u i ior i m n x bear cyrc)1 111 11, trüIi\iates $3 olil) . I

10‘ L error III watw \dtur4t ionq tor e~ii!iplc. u hen the s,itiiratiori equation, Eq. 4.1. is applied.

Second, the depth of ink asiori 15 reflected b j the relativevalue~of the t\ro Inductiori readings. A ratio Rl,,/RID greater than 1.5 indicates quite deep itnwion (d, > 70 in.) and a value less than 1 .I indicates shallow invasion ((1, < 40 in.). These ratios convert to constant separations between the curves on a logarithmic scale so the depth of invasion can be eyeballed easily on the

-1)

logs.

T H I C K BEDS, 8-in t203mmt HOLE

CKiN EFFECT CORRECTED R,, / R, lo0

DIS-EA OR E Q U I V A L E N T

Fig. 4-1 1 Invasion correction chart for Dual Induction-SFL (courtesy Schi urn berger)


Third, the value of R,, derived from the chart, generally 1 .5 to 2 times the SFL reading, is not sufficiently accurate for movable oil calculation betxaiise of approxirnations and uncertaintit35 in the chart. In an!' case inox a- ble o i l calculation with fresh mud is suipect, for reasons given in c1iaptc.r 2 .

D U A L LATEROLOG-R,, L O G S _--_.-_--_______._I_ ___-_

Laterolog qystenis utilize a iniiltiple electrode array to force survey current to t r a ~ e l laterall- across the mud and into the adjacent formation. The advantages that accriie arc the ability to operate in very salt- inrid. excellent bed dcfinition. and independence to neighboring bed resistí\ ities.

'There arc t u o basic types of focused-electrode 1,aterolog arrays. One is the %electrode system, commonly called Guard Ing or LL3. and the other is the 7 tn 9 electrode systvin. with designations such as J L 7 , IL8, L,L,,, and LI+. Both syst~ms operate on much the same pnnciplc, as illustrated i i i Fig. 4-12.

Loleroiop 7

Fig 4-12 Basic 1 aterolog arrays (courtesy ScMumberger)


Cimsidering first the Li,3, survey current i,, is sent out írom the center electrode A and bucking current i, is sent from the guard clectrodes A , and A:, Lvhicli are connected together. The bucking current is adjusted to iiiain- tain zero voltage between A,, and the A,-& pair. There is then no ciiirent flowing iip or down the hole, which means the survey current is forced to flow i n a lateral sheet into the formation. The width of the A,, electrode, commonly 1 to 11 in., determines the vertical resolution of the tool. 'I'he length of the guard electrodes, usually 2.5 to 5 ft, and the proximit!. of the survey current return point determine the depth of penetration. 'I'he longer the giiards and the more remote the return point, the deeper the penetration. Formation resistivity is proportional to Víi, where V is the common electrode potential relative to a far electrode.

In the case of multielectrode systems, illustrated by the L,ater«log 7 , survey current flows from the A, electrode and bucking current flows from the A , and A, electrodes, which are connected together. The bucking current is adjusted to maintain zero voltage across the monitor electrodes M I and h l i ' (which are connected to M, and & I 2 ' , respectively). 'The net result is the same as for the guard system.

THE DLL-MSFL TOOL

Fig. 4-13 is a schematic of this system. The main part of the tool consti- tutei a 9-electrode array that provides deep (LL,) and shallow (LL,) resistivity ciirves. On the bottom of the tool is a pad-mounted, Microspherically Focuced array that provides a flushed zone (RJ resistivity curve. The four- arm linkage that supports the MSFI, pad provides a caliper curve and also centralizes the bottom of the tool. The top of the tool is centered with another centralizer.

Fig. 4-14 shows how the same set of electrodes is used to obtain the deep and shallow curves by using currents at two different frequencies. The deep mcasiirement is made at 35 Hz and the shallow one is made at 280 Hz. The L& achieves deep penetration by having a long electrode array (18 f t ) and returning the current to a surface electrode. With the LL,, current is r~tiirried to a nearby electrode that giws it shallow penetration. Beam width and therefore vertical bed resolution is 24 in. for both curves.

Fig. 4-15 shows the MSFL array. Five rectangular electrodes are mounted on an insulating pad that is forced to ride the side of the hole. Ciirwy ciirrent i, flows from A , and bucking ciirrent i,flows between A, and A, The latter ciirrent is adjusted to rnaintain zero voltage between the monitor electrodes indicated. This forcei the survey current directly into the

c

4 I N


formation, where it bells oiit quickly and rctiirns to a riearb!. electrode. The voltage V between electrode M,, and the inonitor electrodes is inea- si.ired.Resisti~,it!. is proportional to Vi ¡ , , . With this system the MicroSFI, has sufficiently shallow penetration to read flushed-zone resistivit!.. R,,,, directly. even in the presence of miid cakes i i p to 3/4 in. thick.

DEPTH OF INVESTIGATION Fig 4--16 shows the depth of irivestigatiori of the JL,, aiid LL,, ciirves

along Ivith that of the older LL3 and LL7 tools. no\v obsolete. The pseudogeometrical factor. J, is tlie relative weight that the particular array assigns to the invaded zone. The term pveiido- is used because the weight

.-.. c 7

O o m

m

al

L

I

L - .- o I

E, zs, O U al

a

1 .o

0.8

0.6

0.4

0.2

O

I /

1 1 , I

Thick beds //y 8-in hole I R,, = O 1 R,

I

40 60 80 Diameter d,, in.

O 8 20

Fig 4-16 Depths of Investigation of DLI.-Rxo cwves (courtesy Cchlumberger, 'ii SPE-AIME)

tlq)eiids on the relative rcisistivit ics of invaded and rinirivaded zoncs as well as the invasion diameter.

'The invaded and iiriinvaded zones cont r ih te to tlie total resisti\:iiy read b y tlie tciol in accordance ~vitli the lmduct of their resistiviiies and \vcigltts. l ' h n t is

This equation illustrates \vhy the Laterolag is better than tlie Induction 10g in salt miid. Consider the same salt-Iiiiitl sitiiatiori as for the indiiction case: I{,,, .= O.(),j. R,,, = 1, I{, = 10, d, = 65 in. The LL,, geornetricid factor is 0 . G . Cuiisecliiently, the LL,! woiilcl read an appa rmt resistivity

11, = 1 x 0.42 i- 10 (1 - 0.42) == 6 . 2 ohm-m

Thi\ is considerably closer to thc true 1 alue than the Induction reading, wliich would b e less than 3.8 ohin-m. Ne\ ertlieleFs. a siqnificant correction factor is required, which illustrates thc need for a three-curie cornhination.

For the same fresh-riiiid condition5 ay previously cited, R,, = 1.0, R,, = io, R, = 10, d, = 65 in., the Laterolog would gi\e an approximate value

€1, = 20 x 0.212 t 10 (1 - 0.42) = 14.2 ohin-ni

The actual value would he higher because J (at 65 in.) \voiild 1)c grwter than 0.42 \\ it11 the inore retistn c> invaded zorie. The Induction log \\ oiild read 11 nhrn-in under these coriditions; it \\ o d d be the hetter choic$e. 'Thiq illii\ti ate\ that the I,ateroloc tool i\ ad\ vrselv affected l)! a n invaded 7onv inow rp*.i\ti\ v than thc, iindi~tiir1)ed forrii,ition, 1% Iiicli i \ oppmite to the. Iiitiiictiori caw

Vertical Resolution

A s inrntioncd, the vcrtical rtwdiition of íhe and LI,, curve5 i s 2 ft. Adj.went hed cffcct i s c i i i c i l l . i ' No lwd thickness corrections are required.

88 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATtON

CHARACTERISTICS OF THE MicroSFL (MSFL)

The MicroSFL succeeded the Microlaterolog and Proximity log as a device to measure flushed-zone resistivity. The 50 % depths of penetration of the three tools for honiogeneous formations are approximately 1,1, and 3

Therefore, the MSFL has the advantage when invasion is shallow. Bed resolution of the MSFL and other R,, tools isextremely good-on the

order of O in. In fact, there i s so much detail the curve i T often averaged over 2 ft during recording to make it niore compatible with the LL, and I.L, curves.

Borehole (i.e., mud cake) corrections for the MSFL are negligible (< 15 '70) for mud-cake thicknesses between Iá and 34 in. On the other hand, the Microlaterolog requires significant correction for mud cakes greater than in. thick. Since mud-cake thickness is rarely known with any accuracy, the MicroSFL is the preferred curve.

eiy; 80% depths of pe ce tho

G PRESENTATION

Fig. 4-17 is an example of the DLL-MSFL log in salt mud. A CR curve, which caxi be run simultaneously, is presented in Track 1 since the SP is poor in salt mud. These three resistivity curves are recorded in Tracks 2 and 3 on the several-decade logarithmic scale. Normal presentation is LL,, heavy- dashed, LL, Iight-dashed, and MSFL solid. With salt mud the shallowest curve reads lowest resistivity and the deepest Curve reads highest in Lvatcr- bearing zones, which is the reverse of the fresh-mud situation.

Determination of Ri

Correctioii of the L t , to obtain R, i s made by means of Fig. 4-18. The three resistivity curves are read at the point of interest and the chart is entered with the ratio Rr.LdiR,-l ~ on the horizontal scale and R, ,,ill,,, (where R, = HLfIFI) on the vertical scale. From the point of intersection the correction factor R,iR, 1.d is read from the dotted lines trending northeasterly. For level A of Fig. 4-17, RLLd/RLL, = 2.2 and RLLd/R,, = 3, which gives R,iRILd = 1.6. R,isthen6.3 x l.60r10ohm-m. Invasiondiameter, asreadfrom the dashed lines, is 80 in., which is quite large. Notice that the invasion near the bottom of thesection is much smaller because the LL, curve falls closer to the LL, curve. For level B the chart indicates an invasion diameter of only 40 in.

RESISTIVITY LOGS 89

QUICK-LOOK HYDROCARBON INDICATION Fig. 4-19 is an example from the Middle East where the DLL,-R,, tool

was run, not because of salt mud, but because hole size was large, and because in the hydrocarbon-bearing zones (denoted 2 and 3) R, > R,, and H,

o

With the DLL-R,,, combination an approximate water saturation can be estimated from the resistivity logs only. Combining the saturation equations

Fig. 4-1 7 Dual Laterolog-Rxo log in salt m u d (courtesy Schlumberger)


for the undisturbed and fliishd zones, as given by Eqs, 2.S and 2.12

100

00

60

40

30

20

15

I C

8

2 6 U \

- 4 -1

d ? 2

1.5

I

0.8

0.6

0 . 4

0.3

0.2 (


Experience in the industry has indicated that for average oil conditions

s,,, = s,,"? (4.6) (4. $5)

Combining the two equations

S O \ 4 0.6 O 8 1.0 I 5 2

Schlvmbergei

Y

NO hNNWUS NO TRANSlf'tON ZONE USE DATA COA1PECTEC FOR emwt.E EFFECT

I_ - , .- I

RLLd

Fig. 4- 4 8 Invasion correction chart for Dual La?erolog-Rxo (cocirtesy Schlumberger)

(4.7)

Consequently, i f the ratio R,,,/R,\ is knobvn. ivatcir saturation can be estimated directly.

Normally R,,, must be ohtained from mud filtrate measiirenient and H,, needs to be known for the inter\.al in question. Howei-er, even that is not necessary if there is an olwioiis water-hearing interval. In the case of Fig. 4- 19. Zone 1 is water bearing, as indicated by the low value of R,,l,c,. Also no invasion correction is neccssarysince = Rl,L5. Eq. 4 . 5 then gives for this zone, since S,, = S,,, = I

R,,/H, = RJR, = 0.4510.3 = 1.5

Consequently. applying Eq. -1.7 for the center of Zone 2

s, = [(2/650)/1.<5]58 = 0.02

For Zone 3

S,% = [(-i/iLi0j/i.5]'* = 0.12 I . I hese values are only approximate but it is clear the tivo zones i n question are hydrocarbon-saturated.

Use of this technique 1)rovides a little extra time for completion planning ;it the wellsite while the porosity logs are being run. More accurate saturation valiies should be computed using the Archie equation 2.8 as soon as the porosity data is olitained.

The ratio S,I'S,,~ as obtained from Eq. 4.5 is alqo iisefnl as a nioL.able hydrocarbon indicator. If the ratio approaches unity, the tiydrocarbori is immovable. regardless of the absolute values of S, and S,,,: this can happen with very heavy oil. Conversely. i f it is l o ~ v , whatever hydrocarbon exists in the reservoir is certainly producible. Empirical guidelines which havc been estahlished for the Permian Basin of \Vest Texas are"

S,/S,,, l ess than O. 6-hydrocarlion production greatcr than 0.8-\vater production Iwtweeri O , 6 and O . 8.- product ion tcst re< luircd


DUAL LATEROLOG Rxo

OHMS MI/M

LATEROLOG DEEP

LATEROLOG SHALLOW

MICRO S F L

w 1Qo 98)

Fig. 4-19 Dual Laterolog-R,, log in large hole, high resistivity conditions (courtesy Schiumberger)


For Zones 1 and 3 of Fig. 3-19, Eq. 4.5 gives S,IS,, values of 0.04 and O. 18, respectively. Therefore, excellent production would be expected from these zones.

In cases of old wells where resistivity but not porosity logs are available from log libraries, the resistivity ratio method is the only technique available for estimating water saturation. For best results a true R,, curve, from

the SFL, LL-8 or 16 in. Normal curve5 are used in lieu of an R,,log, has been used with success.'b ''

Local experience may dictate an exponent other than 0.2 in Eq. 4.6; it may vary up to 0.5 in the case of gas-bearing or viscous oil-bearing formations where flushing action is poor. Sometimes, rather than using Eq. 4.6, an educated guess of S,, is inserted directly in Eq. 4.5. Values from 0.8 for medium gravity oil to 0.6 for gas or heavy oil are typical.

SUMMARY

mud with medium to low r mud or high formation resi

Modern resistivity combinations are Dual Induction-SFL in fresh ity and Dual Laterolog-R, for salt

QUAL I N D U C T I O N LOGS 0 Run when Rmr > 2Rw and R, < 200 ohm-m. * IL, reading normally close to R1. Correction for invasion usually

* Vertical resolution-4 ft for Induction, I ft for SFL Logging speed-5,000-6,000 ft/hr Should be run with 1 5 in of standoff 10 minimize borehole effects

* Can be recorded in oil-base (nonconductive) mud or in empty hole No shailow curve obtained

0.7 $0 1.0

QUAL LATEROLOGS 0 Run when R,, < 2R, or when R, > 200 ohm-m.

LL, curve should be corrected for invasion to obtain R,, even for wellsite interpretation; correction factors can be a s high as 2.

0 Vertical resolution-2 ft for Laterolog, < I ft for MSFL. Logging speed -5,000-6,000 ft/ h r .

0 Quick-look hydrocarbon saturation and movable oil indication can be obtained.

94 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION Chapter 5

REFERENCES I S. J . Pirson, Ilandbook of \Vcll Log Atial~ysis (Engleivootl Cliffs. KJ: Pren-

tice-iiall Inc.. 19631.

SPL\’I,A. 1979). ? r),\v. Iiiichic, ~ í r í Elcctrica/ ~,og íiitrrprriaiiort, (Goicieii. CO: i).\tT,

liilchie Iiic.. 1979). ‘ M.P. Tixier, R.P. Alger, arid D . H . ‘Tarigii!.. “Ne\$. Dcidopirieritc in IrldLic-

tion and Sunic Logging,” SPE 1300-G. Dallas (Octoher 1 N.A. Schiister. J .P. Radon, arid E.R. Robbiiis. “Application of the

ISF6oiiic Combination Tool to Gulf Coast Formations,” -I‘rciti.sactioris uj Ciilf Coast Association o f Grological Sorietirs (1971).

“R.P. Aiger. \V.P. Riggs. and B.N. Carpenter, “i>uai ~ i i d u c t i o n - ~ . a t e r o ~ o ~ : A N e w Tool for Resistivity Analysis,” SPE íZ3 New Orleaiis (October 1963).

’ H.G. Doll. “The hlicrolog.” Trans. A M E . Vol. 189 (1950).

‘ I H.G. I>oll. “The Microlaternlog ,” J o u r . Pet. Tech. (Jani iary 1953). l o J . Siiaii, 1’. Criinaldi. A. Poupon, arid P. Scwhaite. “The l>i inl Laterolog -

I ‘ P. Soiihaitc, .4. Misk. arid A. I’oiipon. ‘‘€1, Deterinination in the Eastern

H . G . Doll, “Introduction lo Induction Logging,” Jour. Pct. Tech. ( J i i r i r

J.H. Moran and K.S. Kunz, “Basic Theory of Induction Logging,” Gco-

Sclilumberger. Log íntcrp~ctatiori Prítiriples (1972). C . Horst and L . Creagar, “Progress Report on the Intrrprrtation of the Dual

Laterolog---R,,, Tool in the Fermiari Basin.” SPU‘LA Logging Synlpnsirint T r n m ~ r - tion.7 (June 1074).

“h1.P. Tisirr, “Electric Log Analysis in the Rocky hioiintains,” Oil L- Gas Jourtuzl íJiine 23. 1049).

CP\Vl,A. IIonston Chapter. 7’hr ,4r f ~f /trici(wt L o g ,4rili/!/

1 i .G . Doll. “The Laterolog,”Joiir. Pet. Tech. (Noverntier 1951).

R,,, Tool,” SPE 4018. San Antonio [October 1972).

Hemisphere.” SPU.’Z,A Logg ing S?jntpo~iut?i Trans. (June 1075).

1949).

physics (Decemhrr 1962).

12

13

I4

is

Schliinil~erger, I,«g Znfcrprctotion Charts ( L979). 17

POROSITY LOGS eturniiig to the hasic log iriterpretatiori equation

~~ - - R s,, = c \pi,, R t / d (5.1)

we see that porosity is the third and final iiipiit needed to calcrilate \vater saturation. An accurate value of porosity is required since any error i n its valiie will translate to th me percentagc error in water satiiration. ‘The error \vil1 be magnified i n calculating hydrocarbon volume, 4 ~ ( 1 - S,,). For cxarnple, if C$ is too low hy 10 % . S,\ \vil1 he too high by 10 % and liydrocar- bon volume \vil1 be too low by 20% (for a u,ater saturation of 50 ” O ) .

There are three porosity-rneasiiring tools in common use at the present time -- Density. Neutron. and Sonic. il’hl. three when oril!, one valiie is needed? It is hecause all three t o c i l s rcspmd not only to porosity h i t also to the type of rock matrix and to the make-up of fluid filling the pore space - principally Lvhether it contains gas. iVhen the niatrix is known arid pore fluid is al l liquid, onc rneasurenient may suffice. In other cases a l l three measurenients are needcd to sort out the parameters.

THE C U R R E N T TREND IN P O R O S I T Y L O G G I N G

For m a n - years the Sonic was the popular porosity tool. It \vas less sensitive to borehole and miid-cake variations than early Density and Neii- tron tools. It could be run in conibination with the Induction, giving both H, and C$ values for Eq. Fj.1. In shaly sands H, \voiild be atmorinall!r OW and 6 would be almorxnally high, providing a compensating effect such that Eq. 5.1 ~ v o u l d give reasonable xvatctr saturation \dues evcn mdcr these conditions. Porosity values, hnvever. .ir-oiild be optiniist.ic.

In recent ycars the 1)eiisity-Neiitron coinbiilation has heconie the primary source of porosity information. displacing the Sonic. There are several reasons:

Porosity can be deterrriined \I-ithout precise knowledge of rock

There is no need for the compaction correction required wit-h Sonic

Overlay of Derisity-Neutron curves is an excellent gas indicator.

matrix.

porosity.

95

96 ESSENTIALS OF M O D E R N OPEN-HOLE L O G lNTERPRETATlON

0 Transitions from one type of rock matrix to another can often be

0 Shale effects are more evident and can be accounted for more distinguished.

precisely. Consequently, the S

where the hole is very irregular, where secondary porosity is important, or where heavy minerals such as pyrite adversely affect the Density. It is also required when a synthetic seismogram will be generated (from Sonic and Density logs) for depth calibration of seismic sections. This is, in fact, the application for which the Sonic tool was originally designed.

R E C E N T D E V E L O P M E N T S

The porosity tools in common use are the Compensated Density, the Compensated Neutron, and the Compensated Sonic instruments. These

third, in mud-cak

predecessors. They have been in use for the last 10 years.

the Litho-Density, the Dua These tools provide new a The Litho-Density indicates the type of rock matrix - whether sand, limestone, or dolomite, for example, The Dual Porosity neutron gives better gas indication in shaly formations and probably will provide more accurate porosity determination in tight formations. The Long-Spaced Son' IC can yield shear wave velocity, which is important in determining mechanical properties of the formation; it is also the preferred Sonic device when holes are large and altered shales exist.

I n addition, a new and different porosity curve is entering the picture. It is the Electromagnetic Propagation log. In essence it gives water-filled porosity in the formation very close to the borehole. Comparison with total fluid porosity derived from the other porosity curves yields water saturation in the flushed zone without the benefit of resistivity Iogs. Consequently, this promises to be an excellent device for oil detection in freshwater areas where electrical logs are not definitive. It also can provide, along with standard logs, movable oil indication in fresh mud areas.

The principles and responses of these new tools, a well as the current ones, are discussed in the following sections. It appears that the Litho-

In each case, however, there is a new generation of tool coming into we:

rmation about the

¡

POROSITY LOGS 97

Density will eventually displace the Compensated Density, and that the Long-Spaced Sonic and Electromagnetic Propagation log will find good use. However, it is a little early to speculate on the impact of the D u d Porosity Neutron.

COMPENSATED DENSITY AND LITHO-DENSITY

density by measuring the attentuation of gamma rays between a source and t@2tor.' Fig. 5-1 ShQws the arrangement of the Compensated Densit . A Source and two detectors are situated on a pad about 3 ft longthat is forced against thesideof the borehole with a backup arm.

ed continuously by the source (typically 0.66 mev channeled into the formation. There they undergo

multiple collisions with electrons that cause them to lose energy and scatter in all directions - a mechanism called Comptoii srattering. When their energies drop below about O. 1 mev, the gamma rays die by a process called photoelectric absorption. Compton scattering depends only on the electron densit?, of the formation (the riumber of electrons per ec), which is elosel) related to bulk density. This i s the basis of the standard density measurement. On the other hand, photoelectric absorption depends on both electron density and the average atomic number of the material making up the formation. This mechanism is utilized by the Litho-Density tool to indicate rock type.

We may visualize at any instant a gamma ray cloud with a radius of a foot or so surrounding the source. The size of the cloud depends primarily on the formation scattering properties, therefore on electron density. It shrinks and expands as the density varies. The greater the density, the smaller the cloud and vice versa. The population of the cloud, however, which consists mainly of very low-energy gamma rays, depends on the absorption properties of the formation. The greater the absorption coefficient, the smaller the population and vice versa.


Fig 5-1 Source-detector configuration of the compensated Density tool (courtesy Cchlurnberger, tZ SPE-AIME]

THE COMPENSATED DENSITY TOOL

The long-spacing detector. tlie primary one. generates a discrete tdectri- cal pulse for each gamma ray that happeiis t o strikr it. This dctector is situated ncar the edge of the cloiid, so it registers niore piulsei as tliv cloiid expands and fewer as it contracts. It is sliielded (as is the short-spacing detector) from the lo\v-ent:rgy ganiina rays arid is insensitive to absorptioii properties of tlie forniation. The net resii!t is that the pulse rate depends only on eicctron density. T!yically, it \vil1 clecrease cxponc~ntially by a factor of

POROSITY LOGS 99

5-10 (depending on instriiiuentation details) as bulk dcnsity increases from 2.0 to 2.7 gicc. This provides a sensitive measurement of density.'

The function of the short-spacing detector is to compensatc for the effects of resiciual mud cake (not plo\ved awa!. by the pad) and iriiid-filled hole rugosity interposed between the pad and the formation. LL'itli noriiial rniid these provide an easy path for gamma rays to channel upward from source to detector, leading t o erroneously lo\v derisity values if not corrected. First-grneration Density tools with only single detectors suffered from this problem. The short-spacing detector provides a pulse rate that is also inversely proportional to density hii t Ivhich has a shallower depth of invcstigation than the long-spacing detector. It therrfore gives greatt'r w(3iglit to the niricl cake mid hole riigixit!..

h l s e rates froni the two detectors are sent tu the surface and arc combined in a computer iising lal,oratory-deriv(~d "sI>ine-aiid-ribs" response data. sliown in Fig. 5-2. to provide two signals for log presentation. One is the corrected biilk densit!., pi>, obtained by extrapolating the response at a point such as O back to the spine, point P, follo\ving the direction of the ribs. ?'lie other i s tlic correction. Ap. shown by the distance P-Q along the spine, whicli h a s heen added to the basic long-spacing density. point Q , to eliminate the mud cake and rugosity effect.

Esamination of Fig. ,552 reveals t\vo important points. First, A p corrections are positi\-e for normal nonbarite mud cakes but can be negative for Ixirite-loaded mud cakes. Such mud cakes can appear niore dense than adjacent formations. in q-liich rase the density derived from tlie long- spacing rate i s too high and must be corrwted downward. The second point i s that l p corrcctions are adeqiiatc up to about O . 15 gicc but not beyond that point.

Log Presentation

Fig. :3-3 shoLvs a typical Compmsatcd Density presentation. The bii lk tlerisity curve. is recorded over 'J'racks 2 arid 3 on a linear scale from 2.Q. 3.0 gicc. Densities t!,picall!. var!' from about 2.7-2.0 glcc as porosity ixries froni 0-40';ó , The Ap correction w r v e is i n Track 3, Lvith zero at the ceiiter and k O. 25 g/cc at the cistrcnics. 'The correction indicated has already Ilecii applied to tlie p,crirve; it is not nrcessar!. to add or su1)tract it again. The hole dianieter. as nieastircd by tlie l m h i p arm, i s presented in Track 1 . If i i Gamma Ray is rrin siiriiiltaricoi~isl'. u.Iiicli is common. it is also recorded i:i Track 1.

In anal!.zinga Dcnsit!.!og. i t i s liest i n first chseri~c. tlic: &I recording. I t i s a ciiialit!.-contro! ciirvc. In smooth Iiolc i i shoiild be close to tlip zcro linc. a

1 O0 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION

2 9//

SHORT SPACING D E rEcioR COUNTING R A T E

Fig 5-2 Derivation of density and densily correction from short and long- spacing detector rates (couríesy Schiurnberger)

little to the right for normal (nonbarite) mud, and to t!ie left for heavily loaded barite mud. When mud cake or hole rugosity is encountered, the Ap correction will increase. As long as Lip is less than O. 15 gicc, the correction is adequate and the p b curve can be trusted. Above O . 15 glcc the correction is likely to be inadequate and the & curve in error. Corrections are barely adequate in the washed-out shales from 1,852-1,878 ft and

. B’m

200

. . . . . . . . . . . . p . . . . . . í ! : ; L: : ; ..1! : . : : : : : : : , . _ . . . 1.: :.: : ;.: : 1.. ! :.::.:.I L . . - . ....... - . . ~ a--.- - -.- .., . ~ ~ ~ . ~ . . - . . - . . , . . . . L . . . . . . .

Fig. 5-3 Example of a Compensated Density recording (courtesy Cchlumberger]

from 1,924-1,910 ft. The tool cannot compensate for short caves that the pad bridges if they are greater than about ;i in. in depth. Correction will be inadequate and the p,, curve will read too low in normal mud. It also will not read correctly when the pad tilts on entering or leaving a sharp cave. A good example of these effects i s at 1,832-1,834 ft where the low-density

1

102 103

reading is clearlj. due to a short sharp hole \vasliout h i t the & ciir\re sho\r.s almost no correction.

Because of the statistical fluctuations described in Chapter 3. the Den- sity curve will not repeat esactl!.. A\ crages should agrec, hut tile standard deviation between repeat runs \vil1 lie about 0.04 g!cc at high density arid ahout 0.02 gicc at I o ~ v density. Ivitii the nornial averaging time of 2 sec and logging speed of 1,800 ftihr. Nonrepeats may be aggravated by the fact that the tool can ride different sides of the hole on repeat runs. C’l’here viigular porosity exists, formations may not be uniform around the hole.

Electron Density vs Bulk Density; Log Correction

As described. the Density tciol responds to the electron density of the formation. The desired quantity, however, is biilk density. The tnro densities are related by theZiA ratios of the elements making up tlic formation. Z being the atomic number and A the atomic weight of a given cleinerit. For all elements in sedimentary formations except hydrogen, the ZiA ratio is almost constant, varying only froin 0.48 to 0.5. For hydrogen, however, the value is 1 .O. Consequently, the presence of water or oil in formations significantly disturbs the usilal proportionality betwwn electron and bulk density.

The Density tool is calibrated to read bulk density correctly in freshwater-saturated limestone forniations, using test formations of precisely known bulk densities. The result of this calibration is that other formations will read a little incorrectly if their electron densities differ from t h e of limestone-water mixtures of like bulk densities.

The difference between true bulk density. ph, and log-indicated dens it!^. pi, ,e, is shown in Table 5-1 for various substarices of interest. Diffcrcnces are ncgligible for quartz. dolomite, and calcite. However, they are signifirant for beds such as sylvite, halite, e p s u m , and anhydrite. The first tu-o \\.ill record bulk densities about 0. 12 gicc lower than their true densities, and the last two will record about 0.02 gicc higher than true densities. iI.’hen these discrcparicies are seen on a log. they should not be judged faulty tool operation; tliey are the result of the particular calibration choice.

Fig. 5-4 gives the corrections to be applied to log readings to obtain correct bulk densities. Correct.ions are zero or negligible in liquid-saturated limestone, sandstone and dolornite formations. For gas-saturated forina- tions, however, log-indicated densities will be slightly low (such formations having lower electron densities than water-saturated formations of the same bulk densities). The worst case is that of very low-pressure gas or air iri the pores at a porosity of 40 % ; the required correction is about 0.075 gicc at 1.6 gicc, or almost 5 YG . Thc correction becomes proportion at el!^ smallcir as gas

TABLE 5-1 COMPARISON OF ACTUAL AND LOG-INDICATED DENSITIES

Actual Compound Formula Density.

Pb,glcc __ - _ _ ~ Oiartz s l o p 2 654

Dolomite CaCOSMgCO, 2 870 Calcite CaCO3 2 710

Anhydrite COSO, 2 960 Sylviie KCI 1984 Halite NaCl 2 165 Gypsum CaSO,SH20 2 320 Anthracite 1400 coal 1800

coal 1 500

Salt water 200 000 ppm 1 146 Oil níCHd O 850

Biturninous 1 200

Fresh water H2O 1 O00

Methane CHA Pm-th

Log-Indicated Density,

2 648 2 710 2 876 2 977 1863 2 O32 2 351 < 355 1796 1113 1514 1 00 1135 O 850 1335

Piog, g/cc _-

Pmp,h-O 188 1 325 pg-O 188

- Gas c 1 AHA: P9 ____

Source Schlwnberger iog in’terr?ret~tion/Principles

drrisity increases and gas saturation decreases in the formation porcs. I n nornial logging no corrcctions are needed; i h e too1 reads bi i lk drnsit!. di rrct 1y.

Fig. 5 - 6 1 : als« s1ioM.s the correctioris rerliiircd for magnesiiilii and alunii- niini. These points are significant because blocks of these materials are iised ;IS sccoiidary calibratiori standards in field locations.

Depth of Penetration and Vertical Resolution

The 90% depth of investigation of the Compensated Drnsity log is approxiiiiatcl!. 4 in. from the borehole \val1 at mid densities, slightly greater at lowcr densities, arid slightly less at high dcnsities. This nieans the log senses the fliished zone. which contains mud filtrate arid possibly residual h>.drocarbon in the port’s. There is usiially insufficient difference in density bct\veeii water and o i l for the Densit!. to sense residual oil in the fliished 70nc . On the other hand. it can readily senst! residual gas, especiiill!. if porosii!. is high and gas pressurr is lo\v.

1 OA ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION

- - PI,, i gm/cm” -

Fig 5-4 Corrections to obtain true bulk density from log density (courtesy Geoph ysfcs, reprinted by Schiumberger]

The vertical resolution of the tool. if run very sloívly, is approximately 1.5 ft. Formation density i s averaged over thdt interval. Elowever, with the usual averzging time and logging >peed, bed resolution is dbout 3 ft (we Chapter 3 ) .

Borehole Effects

In fluid-filled holes the Density tool response is independent of boreholc size in the range of 6-9 in. This is a result of good shielding or1 the back side of the pad. In larger holes 0.005 gícc should be added to the log reading for each inch of hole diameter above 9 in. for best accuracy.

The Density tool works quite well in empty holes. The response is not the same as that shown in Fig. 5-2, but the appropriate data is utilized in the surface computation of density. Tool response i s independent of hole diameter in the range of 6-9 in.; above that point 0.01 g/cc should be added to the log reading for each additional inch of hole diameter. Mud cake is not

POROSITY LOGS 105

generally a problem in empty holes, which are usually air drilled. However, hole rugosity may be a problem since extremely low-density material (air) is interposed between the pad and the formation. Much less rugosity can be tolerated in air-filled holes than in liquid-filled holes.

G

Porosity is derived from bulk density in a very straightforward manner. For a clean formation with matrix (or grain) density pma, fluid density pr, and porosity 4, the bulk density, pb, is given by the summation of fluid and matrix components

(5.2) Pb = d * Pf + (1 d)pma

Matrix densities in g

andstones, and quartzites = 2.68 for limey sands or sandy limes = 2.71 for limestones

In liyuid-bearing formations fluid density is typically that of the mud filtrate

pi = 1.0 for fresh mud = 1.0 + 0.73 N for salt mud

where N is the sodium chloride concentration in ppm X 10 - ’. Porosity may be derived from Fig. 5-5, which provides a graphical

solution to Eq. 5.3. Bulk density is entered on the bottom scale and porosity is read on the vertical scale for appropriate values of p,,, and pr. As an example, consider the interval 1,899-1,905 ft in Fig. 5-3 where the log density averages 2.29 glcc. Assuming the formation is limestone and the fluid density is 1.0 (fresh-mud filtrate), the derived porosity from Eq. 5.3 or Fig. 5-4 is 24.5 % .

It is more important to know the precise matrix density at low porosity than at high porosity. For example, at pb = 2.6 glcc, derived porosities would be 3% for sand and 6% for limestone. These differ by a factor of 2 and could mean the difference between expecting commercial and noncom- mercial production since a cutoff is often set around 5 % . On the other hand,

106

1

ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION POROSITY L O G S 4 07

Pb, BULK DENSITY, q r n / c c

- Fig. 5-5 Determination of porosity from bulk density (courtesy Schluniberger)

at ph = 2.2 gicc, derived porosities \vould be 27% and 30 'io, which differ only by 10%.

Many logs today have density-derived porosity curves recorded as the log is run. To effect this the logging engineer must insert values of matrix and fluid densities into the surface computer, which is continuously solving Eq. 5.3. Normal choices for matrix drnsity are 2.65 (SS), 2.68, or 2.71 (LS); those for fluid density are 1.0 and 1.1. The logging engineer chooses values generally applicable to the area. It is important that these values be shown on the log heading since the reading must often be corrected to a different matrix value more appropriatc for the particular formation being analyzed. Esamplcs of Density porosity curves, over1 ain with Neiitron porosity, arc given later.

Heavy minerals iri the formation such as pyrite (FeS,) increase the effective matrix density and cause derived porosity to be too low if not taken into account. Occiirrence is not frequent hiit is important i n a few, areas. particularly Alaska and the North Sea."

Effect Of Gas

As described in chapter 1, considerable gas can be left i n the flushed zone of a gas-bearing formation, bypassed by the invading filtrate. The density of the pore fluid can then be considerably less than one. Consequently, in gas- bearing formations there is a dual dilemma if the Density log is the only porosity curve run. First is recognizing that there is gas present, since the curve simply shows a decrease in bulk density that would normally be interpreted as an increase in fluid-filled porosity. Second is determining the correct porosity. It is not a straightforward mattcr. To apply Eq. 5 . 3 , the fluid density, pr, in the zone of investigation must be known. This depends on the water saturation in the invaded zone, SI", the mud filtrate density, P,,,~. and the density, pi,, of the gas in thc pores. That is

Pr = P m l . S," + Pi, (1 - S,J (5 14)

Gas density can be estimated from Fig. 5-6. However, S,,, is not known beforehand. If an R,,, curve is available, then

S."" = c 4 z 2 4 (5 .5 )

If an R,, curve is not available (usually the case with fresh mud), one can

(5.6)

Eqs. 5.3, 5.4, and 5.5 or 5.6 can be solved simultaneorisly or iteratively to give an apparent porosity 4#. Allolving for the electron density effect gives a final porosity

make an assumption siich as

S,,, = s,,''2 = cc t~~,Ti,/+l'

4 = 4,(0.93 -t 0.0 ' ;pr ) (5.7)

As an example, consider the same interval (1,899-1.905 ftj in Fig. 5-3. iissurne it is known to contain gas and that the electric logs gi1.e R, = 0.05 and R, = 40 ohm-m for this interval. From Fig. 5-6, ph = 0.07 gícc. Following the procedure outlined using Eq. 5.6, q5 = 17.5 "/o, This porosity value differs significantly from the 24.5 % foiind on the assumption of 100',70 liqiiid saturation.

This procedure is ci:rnbersome and inacciirate, arid it is rarel!. used. The Density log really nceds outside help to establish matrix t>.pe. identify gas.

I

1 O8 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION 109

I 0.&3* ($3 5s 8 = ~ . s s = > $ e kmci - I- 2.71 -2.3

V*O - 277 - O - S L L

Gas density (grnec)

I o. 1 0 2 0.3 1 I I

2

4

6

a

10

12

14

16

18

I -.I--- __--

Fig. 5-6 Estimation of gas density (from Applied Openhole Log Interpretation, courtesy D. 'VZI Hi I ch i e]

aridsiniplif) porosity determination when gas is present. As weatiall see. the Neutron log fulfills this requirement admiralily .

THE LITHO-DENSITY LOG

The Litho-Density tool (LDT) ' is a third-generation density instrument that provides in addition to the bulk density log, p L , a photoelectric absorption curve, Pe.5 This curve reflects the average atomic number of the formation and is therefore a good indicator of the type of rock matrix. It is helpful in complex lithology interpretation.

'Offered only by Schlumberger at present and designated LDT.

3

Measurement of P, The source-detector arrangement of the LDT tool is basically the same as

that of its predecessor, the FDC (Fig. 5-1). The operation, however, is different. With the LDT, p o and P, measurements are made by energy

shown in Fig. 5-7, which is a plot of the number of detector, as a function of their energy, for three formations having the same bulk density but different volumetric ab$orption indices, U, designated low, medium, and high.

The basic density measurement is made by registeringonly those gamma rays that fall in the high-energy region, designated H. In this range only scattering of the ganima rays is taking place and the number of gamma rays, represented by the area under the curve, depends on electron density only. Conversion of pulse rate to bulk density and correction for mud cake and rugosity is carried out ir1 the same manner as for the Compensated Density tool. Statistical fluctuations in computed density, however, are reduced by a factor of about 2, to the range 0.01 to 0.02 gicc, by utilization of more

that rea n &-

efficient detectors.

Fig. 5-7 Detection windows for the Litho-Density tool [courtesy Schlumberger and SPWLA)


The photoelectric nieasiirement, Fe, i s made by registering those gnmnia rays that fall in the energy \vinclow, S, positioned at very low energy. Iii t h i s region gamnia rays undergo photoelectric absorption as the!. interact ui th the electrons present. The a.bsorption rate dcpends on the product of the absorption coefficient pcr electron. P?. and the electron density. p,,. The pulse rate in the counting ivintlouz therefore responds to a photoelectric absorption index @\ven by

u -= P, + p, (5.8)

The larger the \.aliie of U, the snialler the &e rate and vice versa. LVith suitable calibration, the valiie of ü for any qi,ren formation can br. tleter- mined.

The electron density or* is related to bulk density (:is a result of cali- brating the latter in water-satiirated liniestonc) by the relation

p r = (0,) + 0.1883)/1.0704 (e; . O)

Consequently, from Eqs. 5.8 :ind 5.9

p = 1.0704 . . -. u __ p r > t 0.1883 15.10)

From the two indepeiidcnt riicasiirrwents. IJ and pi,, tlic valtic of P, is determined.

Dependence of P, on Lithology

The parameter I’, rrflccts forination iitholokq because it i s strongi!. dependent on the effecth e atomic nunihcr of the rnediiim abwrhing the ganima r a y . For asirigle element of atomic riuniber Z. P, iq g1vr.ri,111 logging units (barmielectron, with 1 barn = 10 ’’ crn’)), bv

(5. 11)

For a formation containing a number of elements. the effective P,. value is obtaining by siimniing the (ZilO).’.‘ values, after weighting each by its relative electron densit’. in the niistiire. Table <5-2, column 2, gives effective P, values for common sedimentary materials. P , values for quartz, calcite, and dolomite are quite distinct. Anhydrite and calcite have similar P, \.aliies but are well separated in density (column 3). Xlinerals siich as sicleritc

-- * p e ~ i a s units of 3.0 x, IO’” eiictronsicc.

POROSITY LOGS 1 4 1

(FeCO,) arid pyrite (FeS?) hive coiisiderably higher P, values by virtue of the higher atoniic nuinbcr of iron (Z =r 26). Barite has a11 extremely Iiigh P,. \ , d u t , ; Z = 56 for hariiim. The concecliienct% are discussed b c h v .

Fig. 5-8 s1icin.s P, values for lirnestoiic, dolornitc, and sandstone formations of 0-35 % porosity n.ith I’ores coiitainingeither fresh water or methane of density O . I gicc. Note that regard1t.s of porosity or type of fluid. t h e P, \.slues for the three typw of rock are well separated. Consequently, when only one matris type is present in a formation. the I’,~ curve \vil1 iiriaiii1)igii- ously distinguish it.

Depth of Penetration and Vertical Resolution IIeptli of penctration and vertical rwiliition for the oh incasurcnlent

shoiild be essentially t h sanie as for t.he (hnpensated IJensity. Sirnilar valiies \vould be expected for the P, iiicasiirment, though no verificatioii has Leen piiblished yet.

Borehole Effects For thepb curve, borehole effects are inuch the same as for the Comperi-

sated Density tool. Mud cake and riigosity corrections are made via the

Pe --__t

1 2 3 4 5 6

í!> t

c

0 5 0 4 0 3 0.2 O 1 0

- _ _ ~ __ Fig 5-8 Photoelectric absorptiov coefficient as a function of porosity and fluid type (courtesy Schhmberger and sPWLI\)


TABLE 5-2 VALUES OF PHOTOELECTRIC ABSORPTION COEFFlClENl PER ELECTRON, P,, AND PER CC, U,

FOR VARIOUS SUBSTANCES

U --_I_

p* - Sp. gr. PbOOQl

1 8 1 2.65 2 64 4 78 Qiiartz CQlClte Dolowte Anhydrite Halite Siaorite Pyrite Barite% Water (fresh) Water (100K ppm NaCI) Water (200K ppm NaCI)

5 O8 2 71

5 05 2 96 4 65 2 47

44 7 3 94 17 O 5 o0

267 A 48 o 3513 1 O0

o 734 1 06

1 1 2 1 1 2

3 14 i a 7 2 74 2 88 2 98 2 OA 3 89 4 99 4 o9 1 O0

1 o5

1 1 :

13 8

44 9

55 9

3 00

3 68

a2 I 1365

O 398

o 850

136

spine-and-ribs method. Corrections are ad ong as Ap i,c, Iejs than 0.15 gicc. For theP,.cirrvethesituation isle inforniation hac been published on whether the U measurement is Compensated for mud cake and rugosity. These data are needed Barite-loaded drillinq mud presents a difficult problem for the Litho-Demity log. IE barite-loaded mud or mtid cake intrudes between the pad and the formation. as is almost incvitable. the very high I', value of barite swamps the formation value, rericl~ring the P, curve iiseless. This severely liniit\ the iitilit? of thc LDT in arms where barite mud is used. LL'cighting the mud with iron cornpotmdq 11.odd cause much less effect on the P, cur\ e .

Litho-Density E x a m p l e

A I? )-Dens in sim , with Gamma Ra!. and Compensated Neutron is shown in Fig. 5-9. The o,, curve, converted to porosity using a limestone matrix density ( 2 . i l g/cc). i s recorded in Track 3 (solid). 05,erlain (dashed) is the Neutron porosity curve. The I><, ct1n.e is recorded in Track 2. Gamma Ray and Caliper logs are in Track 1. On the w t o m of the example. the P, ranges of sandstone, dolomite, and limestone are marked for the 0-13 % porosity interval of intereqt. Clearly, the forma- .ion at level A is limestone and that at level C i.; almost pure sandstone. At

a -5, \o C c C d conkena bc>c*tcr, nt9 &cc; .ca\ocQj fia

DIAY IN i M W E

- r-7-r-r-r-r-r-r

I

Fig. 5-9 Example of a Litho-Density log (courtesy Schlumberger)

114 ESSENTIALS OF M O D E R N OP€N-HOLE LOG INTERPRETATION

level B, ho\\ ,e\w, the matrix cannot be resolved frori-i the P, ciirve. It c,oiiict tieeitlier a mixtureof about 50% liniestorie arid5O<~ doiornitcor aI>out T O ( < , limestone arid 30 % sandstont. It coiild e\'cii be a conibinatiori nf all tIirc1.e.

LITHOLOGY INTERPRETATION WITH P,-P, CURVES If only two matrix minerals are present, the voluriietric frartioris of cacti,

along with the porosity. can he derived from the coriibinatic~n of p i , arid €',values.".' Fig. 5-10 shows the applicablt~ chart. The log va1ut.s of pi, aiid I', areentered m t l a point of intersectiori i s íoiirid. Selecting the aIrnost-\,ertical

C . Calcite D - Dlilornite Q - Quartz

I O i

6 5 4 3 2 I _---.

Fig. 5-10 Derivation of porosity and lithology for a two-component matrix (after Gardner and Dumanoir, Schlumberger and SPWLA)

POROSITY LOGS 115

stems representing ihr ~ M T J minerals knou.n to l ie present, porosity is tlediiced 1)y iiitcrpoluting bct\\.cen the eqiiiporosiity lines (at 5 porosity-unit intervals) joining the t\vo >terns. Xf atrix fraction is represerited by the dis- lilacwnerit of the p i r i t along the ecluiporosit! line joining the s t t m s .

A s an example, considrr point I3 of F i g . 5-9 and assiiriie the matrix is a calcite-cloloniite niixturc. The log valuc, of p i , is 2.68, corresponding to linie.storie-deiivec1 porosity of 2 Y; (Fig. 5--5) , and I', is 3.0. Using the calcite- dolornite lines, porosity is found to IJ~. ci ?" arid tlie matrix is a 5 0 3 0 mixture. If, ho\vever, a mixture of calcite arid sandstone is assumed, porosity i s 0'70 and the matrix is 24 calcite arid ',i sandstone. Additional information is needed to resolve the question. Utilizing t tie Keutron curve shows the actual porosity is 2 % arid the matrix is about 60% calcite, 20% sandstone, and 20 % tfoloinite. For this piurpose the ~~olurnetric absorption coefficient, U, is iised. \'allies of U are listed iri Tüblf.: 5-2. TIE procedure is described in chapter 6 .

________ ______ __.___- p_______-.

COMPENSATED NEUTRON AND DUAL POROSITY NEUTRON LOGS

____I__________~ __-

In its simplest form a Neutron tool is illustrated in Fig. 5-11. Fast neutrons ( - rj niev) are continuously emitted by the neutron source and travel out in all directioris into the formation. As they progress, they are slowed or moderated by collisions n i th nuclei in their path. When they reach very low or thermal energies (-0.025 ev), they zigzag or diffuse aimlessly until they are absorbed or captured by the nuclei present.8

The element most effecti\re in slimring the neutrons is hydrogen because a neutron and hydrogen nucleus have the same mass. In a direct collision the ritvitron will t r a d e r all of its energy to the hydrogen nucleus and stop dead, as in a head-on billiard hall collision. On tlie other hand, other nuclei conmion to elements iri sedimentary formations, such as those of silicon, calc;iiini, carbon, and oxygen, are much more massive than neutrons. They are effective in scattering neutrons into different directions but absorb only a small fraction of the neutron e n e r a even in a direct collision. Their effect on the neutron slowing-down process is much smaller than that of hydIo- gen, though not negligible.


'J

4 1 I

SMALL HYDROGEN CONI

HlGH COUNT RA1

URGE HYOROGEN CONC

LOW COUNT RATE LOW mnosii nicn rnnosnv

Fig. 5- I I Single detector Neutron tool in borehole environment [courtesy Welex)

The net result is that one can visualize at any inst surrounding the source, extending a maximum of about 2 ft. As the hydrogen content of the formation varies, the size of the cloud expands and shrinks. The greater the hydrogen content, the smaller the cloud and vice versa. The population of the cloud. which consists mainly of thermal neutrons, i s dependent on the absorbing qualities of the formation for such neutrons. The population at any instant is such that the number of thermal neutrons being absorbed each second equals the rate of fast neutron emission from the source. Consequently, the greater the absorption coefficient of the formation, the smaller the population and vice versa.

Situated toward the edge of the cloud is a detector that may be one of three types: a thermal neutron detector sensing the density of the lowest-energy neutrons in its vicinity, an epithermal detector sensing the density of neutrons just above thermal energy, or a capture gamma ray detector sensitive to the gamma rays produced by absorption of thermal neutrons in its vicinity. Regardless of the type, the pulse rate registered by the detector increases when the cloud expands (less hydrogen) and decreases urhen it contracts (more hydrogen). Pulse rate therefore varies inversely with porosity, since all of the hydrogen (in clean formations) is contained in the pore fluid.

POROSITY L O G S 417

NEUTRON TOOL EVOLUTION

First-generation Neutron tools, used extensively in the 1950s, wereof the

of the two. These tools were very sensitive to borehole parameters, and conversion of log response to porosity was subject to considerable error. They have long since been superseded for open-hole logging, but a limited number still are used for correlation logging in cased hole.

The second-generation tool popular in the 1960s was the Sidewall Neu- tron,' a pad device very similar in configuration to the Density sonde. * This tool was much less sensitive to borehole parameters than its predecessor. '41~0 it was insensitive to thermal neutron absorbers by virtue of using epithermal detection. However, depth of investigation was reduced, which

gosity effects and prohibited

filled holes, where it has better porosity response than other types of Neu- tron devices.

The third-generatio nsated Neutron introduced about 1970. is the current Sta es a pair of neutron detectors (thermal) instead of a single one. This provides definite advantages.

A fourth-generation took, the Dual Porosity Neutron, is just being introduced. It uses a pair of thermal neutron detectors on one side of the source and a pair of epithermal detectors on the other. Each set provides a porosity curve. This tool will be discussed later.

THE COMPENSATED NEUTRON

The principle of the Compensated Neutron tool* ' is shown in Fig. 5-12. A fast neutron source is located near the bottom of the tool, and two thermal neutron detectors are spaced 1-2 ft above it. The ratio of the pulse rates from the near and far detectors, N,/N,, is measured and related to formation porosity. It has been proven, theoretically and experimentally, that ratio

'Designated SNP by Schiumherger, SWN by Dresser-Atlas and Welex, and SNL by Gearhart. "Designated CNL by Schlumbcrger, CNLog by Dresser-Atlas, DSN by Welex, and CNS by Gearhart.

kli:

POROSITY L O G S 118 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION

measurement significantly reduces borehole effects and incream tlie depth of penetration of the toul relatiw to a single d(1tcietor measurment. " At tlic same time rietitron ahsorption effects are reduced, though not eliniinaied.

'I'lie actual configuration of the logging instrunient is stio\í~i in Fig. 5- 13. Tlic. Lvhole tool is decentralized by means of a spring. and the ljackside of tlie source-detector arra). facing tlie mud column is shielded as niuch as possit)le to minimize borehole effects.

The relationship between ratio and porosity for the Schluinbergcr CKL, as dcterniined by measurements in laboratory formations is shown i n Fig. Ti-14. The ratio increases \vith porosity because the thermal neutron density falls off more sharply \vith distance from the source as purosit!. iricreaws, even though both counting rates decrease. (Ratio-porosity relationships diifer among tool mociels, depending on detector size, placerriwt, and shielding. )

Fig. 5-12 Dual detector Neutron tool in a borehole environment ( ~ 0 ~ r t t . s ~ Welex)

119

-- Fig, 5-13 Sketch of a CNi. tool (courtesy Schlumberger, O SPE)


r 4

7’/a in. borehole 1 1

asota dolornite rrected for impurities

1$ 1

I O I I 1 I

20 30 40 50 Porosity, %

O 10

Fig. 5-14 CNL response in sandstone. limestone, a n d dolomtte formatrons (courtesy Schlurnberger, (C; SPE-AIME)

While the ratio depends primarily on porosity, there is also a signficant dependence on lithology because the matrix contributes some to the slow- down and capture of the neutrons. It is clear that to derive porosity from the ratio with any accuracy, the lithology must be known.

Porosity Equivalence

Examination of Fig. 5-14 leads to a useful concept, that of equivalent porositicni. Equivalent porosities are obtained by reading the dolomite, limestone, and sandstone porosítíes corresponding to a given ratio. For example, at a ratio of 2.0 we read 8% porosity for dolomite, 15% for limestone, and 19.5 % for sandstone. These are equivalent porosities. Loosely speaking, neutrons slowing down and thermalizing cannot tell whether they are in one or the other of these equivalent formations.

Plotting porosity equivalents obtained at different ratios as a function of the limestone porosity corresponding to a given ratio leads to the porosity-

POROSITY LOGS 121

equivalence chart of Fig. 5-15 (dashed lines) for the CNL. Equivalent porosities are read vertically. For example a limestone porosity of 14 % is equivalent to a dolomite porosity of 7 % or a sandstone porosity of 18 % (liiw A

Porosity equivalents for the Sidewall Neutron (SNP) are also shown on Fig. 5-15 (solid lines). Matrix effects are less than for the CNL because epithermal detectíon eliminates neutron absorption effects that contribute partially to the lithology differences. Because of this, some operators still prefer the Sidewall Neutron over the Compensated Neutron at low porosities. In particular there is some uncertainty in the CNL response to very low- porosity dolomites apparently because of thermal neutron absorbers that are sometimes present in these forniations in trace

When a Compensated Neutron log is run, the ratio is not recorded. Rather, the ratio is transformed to porosity, on the basis of laboratory data

30 X .- L

t U

m o U c

I

.-

.- 20 o

2

- c .- 03

O a

2 ti

io

O

~~

Fig 5- 15 Neutron porosity equivalence curves (courtesy Schlumberger)


such as that uf Fig. 5-11, iii a surfdce computer arid a porosity curi’e ic

recorded. To effect the transformation, the logging engineer must input to the computer which inatrix to use. He has a choice of limestone (IS) or sandstone (SS). The value chosen is the one most appropriate for the area and is shown on the log heading. It is uaually left constant over the whole log even though the matrix may vary in intervals.

Depth of investigation and Vertical Resolution

Fig. 5-16 shows the depth of investigation of the Compensated Neutruri (CNL) tool in open hole at 22% porosity. For comparison those of the Sidewall Neutron (SNP) and Compensated Density (FDC) tools are also shown. None of the tools penetrates very deeply; but of the three the CNI, has the greatest depth of investigation, It obtains 90% of its response from the first 10 in. of formation compared to 7 in. for the SNP and 4 in. for the FDC. Just as significant in terms of suppressing mud cake and rugosity effects is that the CNL receives only about 3 % of its response from the first inch of formation compared to about G 7 0 for the SNP and 17 % for the FDC.

For the Neutron tools, depths of investigation will decrease slightly at higher porosites and increase somewhat at lower porosities. The reverse is true for the Density t00l.l~

. - Depth saturated

Inches from borehole wall - Fig. 5-16 Depth of investigation of Neutron and Density tools (courtesy Schlumberger, O SPE-AIME)


Vertical resoliitiori of the CNL tool, if r u n very slo\vly, i s approximately 15 in. However, with the iisual 2-sec averaging time and 1,800-ftilir logging speed, it is approsiinately 3 ft. Statistical fliictiiations \vi11 average ahorit 1 porosity unit at very low porosities and 3 pu at high porosities (opposite to the Densit!. belia\lior). Chrisequently , sharp pcak values sliould not be read; averages over about 3 ft should be taken.

Log Presentation

The Conipensatcd Neutron is rarely run by itself because of substantial niatrix arid clay effects. It is normall!. run in combination with the Conipen- sated Density and Gamnia n a y in the configiiration shown i n fig. 5-11. The Neutron is positioned above the Density with its backup spring lined up with that of the Derisity so hoth are forcing the array against the same side of the hole.

Fig. 5-18 shonrs the normal presentation of curves obtained with the aforementioned combination. Garnnia Ray, Caliper, arid bit size are re- curded i n Track 1, and Neutron and Densit’ porosities are recorded in Tracks 2 and 3 with the Neutron curve dashed and the Density solid. (In the Eastern Hemisphere the Density curve is often left on a density scale rather than transformed to porosity.) In this example the engineer chose to record porosity on the assumption of a limestone (LS) matrix.

Consider how the Neutron Lvoidd have to be interpreted if run alone. The Gamma Ray shows the whole section to be clean, so clay effect is not a concern. Also the hole is smooth arid close to gauge, so environmental effects are mininial.

The interval from 15,332-15,336 f i reads a porosity value of 14% . If the matrix is irideed limestone, this is the correct \ d u e . Ií, however, the matrix is dolomite, porosity is 7 ‘‘4, ; if sandstone, porosity is 18 %J , as shown by lines A of Fig. 5-15. (Kote that if the log heading had shown SS, the porosities would be read as 11‘X if sandstone, 10 7, if limestone, and 3.5 70 if dolomite, asindicated byliriesB of Fig. 5-15.) OLviously theNeutron isnot much help unless lithology is known.

The same uncertaint) exists if Density alone is considered. For the same interval Ilerisity porosity reads 2 % if the matrix is limestone. Referring to Fig. 5-5 and assuming fluid density of 1 .O g!cc, porosity would be 10.5% if the matrix is dolomite. Sandstone is ruled out since the indicated bulk density, 2.68 glcc, is greater than the matrix density of sand, 2.65 glcc.

By combining Neutron and Density interpretation, the uncertainties of lithology can largely be circumvented. First, however, environmental effects and the effect of gas on the Neutron log need to be considered.

I-

DE N S ITY

I

. \

. . -

7

\’ .? ii

1 7


Mud W l V 3 Ib/gol

Soltniv ( t i h i M kppm b l n i t y l m 15Okppm Stondoff O

r 175 i

LhpL l I x x ) I +

us u ally srii all . €io u ~ \ , e i . , on final i 11 terpre t a t ioi i - 1) art icul ar 1 y one 111 ad c' for reserve calculations- the corrections should he included.

Fig 5-1 9 embodies the riecessar)~ corrections aiid is largely self-esplaiia- tory. Except for nomograph A, correction is made b y drawing a line at t tie

prosit) , of interest do\vn through the other nomographs, 13 through H, arid estimating each correction independently rising the guidelines furnished. The correctioris-some positive, sonic negative--are then added algebrai- cally to obtain the net correction in porosity units, which is added to or subtracted from the log reading.

The largest correction is for borehole diameter. This is made by irieaiis of riomograph A. However, when the Neutron is run lvith the Densit),, the Caliper reading of the latter is used to make the borehole correction aiito- matically while the log is recorded. In this case the Caliper reads borehole diameter smaller than it actually is by the amount of miid cake thickness (because it is assumed the backup shoe cuts though the mud cake but the pad does not). Consequently, the automatic borehole correction is a little too small and must be increased (reducing indicated porosity) before applying the remaining corrections, as illustrated in Example 1 of Fig. 5-19.

Example 2 of Fig. 5-19 illustrates the case when a CNL log is not Caliper-corrected during the run but a separate Caliper log is available. The difference between the actual hole size at a depth of interest and the assumed hole size set into the surface computer is entered in nomograph A. The mud cake correction is ignored and other corrections are applied :IS usual.

Standoff effects can be important, especially at low porosity. 1f;hile the CNL has a decentralizing spring, the tool is usiially run i n a combiii,ition string that extends aboiit 1,5 ft abo1.e the source-detector array and anywhere from 20-40 ft below it. This means that the tool \vil1 bridge many hole washouts, effectildy creating local standoff.

Teniperature and pressure corrections can also Le significant, but fortunately they offset one another. Combined correction is significant onl), i n high-pressure zones \vith low temperature or in the reverse situation.

Net corrections of 1-2 porosity units, illustrated by Esample 1 of Fig. 5-19, are typical. This is the reason they can be ignored on a first pass, escept in cases where porosit!, hovers around cutoff values of about 5 % . Find iorrcclioni DL shown

516 = - 4 O + O - I 5 - 0 7 + 0 - l 5 t 3 3 = 4 ( m = 3 0 - 4 0 = 2 6 p u

itkip Block 8)

___ Gas Effect on the Neutron L0gI5

Replacement of liquid by gas in the pore space of a rock lowers the hydrogen density of the pore fluid. (Lower gas density predominates over

POROSITY LOGS 127

XAMPLf 1. l u l o . Caliper ( i n redl

GIVEN

I28 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION

greater hydrogen fraction in the fluid.) As a result the Neutron curve, calibrated for liquid-filled porosity, indicates abnormally low porosity. The effect can be large. As an example, consider the zone from 1,884 ft to 1,922 ft

approximation, about 2/3 of the pore space in the invaded zone is filled with gas (assumed at low pressure) and % is filled with liquid. In reality, there is somewhat more liquid and less gas than that.

If the Neutron is the only porosity log run in potentially gas-bearing zones, there is the same dual dilemma as with the Density. First is to recognize the presence of gas, since it will appear on the log simply as lower porosity; if there are other low-porosity zones, as in Fig. 5-20, the gas will not stand out. Second is to derive the correct porosity because the gas

ecause the excavation efject must ned as the difference, in porosity a gas-bearing formation and that

formation having the same hydrogen con- because it will contain less rock vel a little further. For example,

a 30 % -porosity formation with 50 % water and 50 % air in the pores would not read a porosity of 15 % , as might be expected, but a porosity of 9 % . The excavation effect wouid be 6 pu, which is not negligible. This difference is caused by the air space not being replaced by rock matrix.

An iterative procedure can be followed io obtain porosity in suspected gas-bearingzones. It is similar to that described for the Density log, utilizing electric logs to provide values of C,, (Eqs, 5 .5 or 5.6) and applying the excavation correction at every iteration. However, this procedure is exceed- ingly cumbersonie and, a. with lithology determination, can be replaced with simultaneous Density-Neutron interpretation.

C O M B I N E D DENSITY-NEUTRON INTERPRETATION

I I

I I I

I I I

i ,

I

POROSITY L O G S 129 d =

Vastly improved and simplified log analysis is achieved when Density and Neutron interpretation are combined. Fig. 5-21 shows the crossplot chart obtained when Density response is plotted against Neutron porosity. It embodies the information in both Figs. 5-5 and 5-15.

Caliper diam in inches Porosity index (59) Llmeclonamatrix Porosity index (59) Llmeclonamatrix

Compensated tormation density porosity

Fig. 5-20 Neutron-Density log through a gas-bearing interval (courtesy Schlumberclerl


The ke). feature of this chart is that the equiporosity lines that join points of like porosity on the three matrix curves are virtually straight. This means that p o r o s i t k c m be read without precise krtowledgc of lithology. For example, the 15,332-15,336 ft interval of Fig. 5-18 where @,, = 14% and +,, = 2% gives apoint of intersection A on Fig. 5-21. Interpolating between the + =5 arid 10% lines yields a porosity of 8 % . Further, the matrix is

2 0 -

2 I -

2 2 -

2 3 - E

= 2 4 - >-

$ 2 5 - W a

5 2 6 - 3

2 2 7 -

\ oi

t- -

m

F D c*- CN L* Pf = 1.0 oSULFUR

'SALT

LAN>üElNiTE , P O L Y H A L I T E I

1-5 D 8"

( + C N L ) c o r NEUTRON POROSITY INDEX, p u

(APPARENT LIMESTONE POROSITY)

Fig 5-21 FDC-CNL crossplot for porosity and lithology determination in fresh-mud conditions (courtesy Cchlumberger)


indicated as mostly dolomite with sonle liniestone (probable) or some saiid- stone (less probable) or perhaps some of both; however, it could not be a mixture of liniestone and sandstone.

Note that the bottom and right-hand scales of Fig. 5-21 can only be used if the log has been recorded on a limestone matrix. Iiad the same log been recorded on a sandstone scale, the intersection of 4" = 14 % and +d = 2 % urould be found by starting at those porosities on the sandstone line and projecting vertically and horizontally, respectively, to the point of intersection B. Porosity would still be determined as 8 "io, but lithology would be deduced as much less dolomite and more limestone or sandstone. To avoid confusion it is good practice to disregard the bottom and right-hand scales of the chart and use only the porosities on the limestone or sandstone curves (whichever the log heading indicates) as starting points.

Liquid-Filled Formations All liquid-filled porosity points of usual lithology will fall inside the

region bounded by the sandstone and dolomite lines. For this region a good approximation to the true porosity is the averageof theDensity arid Neutron values

+ = ( 4 d f 4,,)/2 (5.12)

This is a most important result. Effective porosities can be eyeballed from the log as those values halfway between Density and Neutron curves for clean, liquid-filled formations. (If the Density porosity goes negative, its value should be taken as zero for this purpose.)

Looking again at Fig. 5-18 arid using Fig. 5-21, we can deduce that the interval from 15,250 to 15,300 ft is primarily anhydrite of zero porosity and that at 15,381 to 15,385 ft, where the two curves agree, is limestone of 1.5 70 porosity. The log illustrates very nicely anhydrite, dolomite, and liniestoiie signatures seen in tight carbonates.

Gas-Bearing Formations As previously explained, replacement of liquid by gas in the pore space of

rock causes both bulk density and hydrogen content to decrease. 'The Den- sity will show higher porosity and the Neutron lower. This gives rise to the well-known crossouer effect on Neutron-Density logs. Norrnally, the Neu- tron reads somewhat higher porosity than Density due to dolomitization and to clay effects. When the Neutron crosses over and reads lower porosity than the Density, i t is an infallible indicator of gas (except for the qualifica- ticm given below), This is a very popular feature of the Density-Neutron combinatioii.


Fig. 5-20 is an excellent example. Throughout the interval shown, the Neutron reads equal to (within statistics) or higher than the Density except in the 1,884 to 1,922 ft zone where there is a marked crossover. This interval is clearly gas bearing

the c points to shift northwesterly and in many cases will cause them to fall above the sandstone line. For the interval 1,900 to 1,905 ft of Fig. 5-20 where @,, = 6 and & = 24.5%, the point of intersection is at C. To find the porosity, the point is shifted back to the assumed litholoL9 line (in this case limestone) in a direction parallel to the gas correction line indicated. in this case the porosity is found as 17.5% (point D).

A good approximation to the true porosity in gas-bearing zones is

$J = d(dd* +4,”)12 (5.13) In the case illustrated this formula gives 4 = 18 % , which agrees with the

the density and composition of the gas, which is the reason the line is labeled “approximate.”

i n situations where the depth of invasion is in the range 4-6 in., the Density will respond only to the invaded zone, u hereas the Neutron will “see” well into the noninvaded zone. In gas-bearing intervals the invaded zone will have lower gas saturation than the noninvaded region, perhaps by a factor of two. The appropriate gas correction line then becomes significantly closer to horizontal. Use of the indicated line results in iinderesti- matingporosities. In the extreme situation of 4-6 in. of invasion and practically no gas in the invaded zone, the correction line is virtually horizontal.

Precise correction requires running an R,, log and combining deri\.ed S,,, d,, and S, values with applicable depth of investigation data for the Density and Neutron.” This normally is not done but could lead to improved iriter- pretation in gas-bearing zones.

False Gas indication

There is one circumstance where a false indication of gas can be obtained from Neutron-Density crossover. This is the sitiiation where the porosit).


curves are recorded on limestone matrix but the lithology is actually sandstone. Referring to Fig. 5-21, a sandstone of porosity 10 7’0 -if recorded on limestone matrix-will show up as a Neutron porosity of 6% (point E) and a

Fig. 5-22 is an example. The Cotton Valley formation illustrated is primarily sandstone, but the log has been recorded on limestone matrix. The whole interval appears gas bearing but is not. To correct the curves to sandstone matrix, 3.5 pu should be added to the Neutron and 3.5 pu subtracted from the Density (Fig. 5-21). The curves then overlay except in a few thin beds that appear to be gas bearing. However, even in these intervals much of the residual curve separation could result from hole erilargenient (short cave) effects on the Density log.

The clue to false gas logs is a 6-7 pu constant difference between the hen this is seen, matrix effect should be suspected,

By the same token, gas can be missed in the case where the log i s recorded on sandstone matrix and a low-porosity, gas-bearing limestone is penetrated. The gas shift might be occurring; but if it i s less than 6-7 pu, it would not cause the Neutron to cross over the Density. Gas is also niore difficult to see with a dolomite matrix. Regardless of limestone or sandstone recording, dolomite causes the Neutron to read much higher than Density (a5 shown by Fig. 5-18). 1% hieh can suppress crossover in gas-bearing zones. In cases such as this, it is important to have independent determination of lithology.

The Litho-Density log can provide the necessary lithology information. Fig. 5-23 i s an example from the hfiddle East where the Litho-Derisity- Neutron combination was run through a series of gas-bearing limestones and dolomites. Scales are such that the Neutron and Density curves should o\ erlay (approximately) in liquid-filled limestones. For some intervals large crossovers occiir, and in others the curves agree. Without a P, ciirve these could be interpreted as gas- and liquid-bearing limestones, rebpectively.

The P, curve, however, s h o w the true story. These zones with large crosovers are limestones (Pe = 5), clearly gas-bearing. Thosr with no cross- o\ er are doloniites (Pr = 3). Plotting the latter points on Fig. 5-21 shows the dolomites to be gas-bearing also. Thus, the Lvhole interval, with the exception of a few tight streaks, is gas productive. This illustrates perhaps the prime use of the Litho-Density-Neutron combination-the separation of gas and matrix effects in tight carbonates.

o

._

f

136 ESSENTIALS OF MODERN OPEN-HOLE LOG INTERPRETATION

THE DUAL POROSITY COMPENSATED NEUTRON LOG'*

All Neutron tools, regardless of type, respond to hydrogen bound in lattice of clays. Porosities in shales typically read 30-45 70, dependingon the am the the over

The problem is aggravated in the Compensated Neutron by virtiie of the fact that thermal neutron detection makes the tool sensitive to trace amounts of strong absorbers such as boron and gadolinium in formations. Such absorbers, if present, are typically found in clays. They may contrih- Ute as much as 15 porosity units to the porosity reading in shale, though typically the effect may be more like 5 pu.

Principie of the Tool

er is introducing a

hermal neutron detec- e source are two sity log insensitive

d by the spine-and- ity) to give porosity

corrected for mud cake and rugosity, as illustrated in Fig. 5-25. This method provides better environmental correction than the ratio method and also results in a A4 correction curve that can be used for quality eon- trol. l9

G. Fig. 5-24 shows

Applications of the log

The improved gas indication in shaly formations with epithermal detection i s shown in Fig. 5-26. In the upper zone-marked A-which is a shaiy sand, the epithermal Neutron crosses over and shows appreciably less porosity than the Density-clearly indicating gas-whereas the thermal Neutron shows no crossover. The quick-look appeal is obvious.

Another application where the Dual Porosity Neutron should prove advantageous is in very low-porosity carbonates or tight gas sands. Matrix effect with epithermal detection should be considerably less than for thermal detection, in fact similar to that shown on Fig. 5-15 for the SNP. At porosities around 5 % where the effect is critical, porosity determination should be much improved with the Dual Porosity CNL.

I

Fig. 5-24 Configuration of the Dual-Porosity Compensated-Neutron Schl

Fig. 5-25 Processing CNT-G detector response to eliminate standoff effects (courtesy Schlumberger, O SPE-AIME)

EPITHERMAL NEUTRC EFFECT OF STAND-O

LIMESTONE 8" i


Thermal detection

Epittierinal detection

! .O P m a 3 . c

1 GR i 5 c

Fig. 5-26 Dual Neutron porosity comparison in a shaly gas zone (courtesy Schlumberger, O SPE-AIME)


~ ~

COMPENSATED SONIC AND LONG-SPACED SONIC LOGS

__ ~

.I Sonic logging tool iiicasur<Jh t l i e \ elocit!. oí S O L I ~ ~ iri forriiatioiis p r n r - tratecl b!. tlir tiorehole. Tile priiiciple is illustrated in Fig. 5-27. A transmitter and t\vo recci\,ers arc positioiieti on a sonde \viti1 a rpcirig (typicall! )

of 3 f t l i e tuwn transniitter and near rt.cei\.er and ;I s p i i r ~ of 2 ft t)et\vct.ii receivers.

T

R, Fit,

U,,,

Sonde body

2

Transmitter signal

b

Time

9 :: 40 psec

Receiver signals

-- Fig 5-27 Basic Conic logging system (courtesy Schlumberger, IC SPE)


When a pulse of current or voltage is applied to the transmitter, it generates a short oscillatory pressure pulse at about 25 kHz frequency in the mud. This initiatessix different waves traveling up and down the hole: two refracted waves via the formation (compressional and shear), two direct

borehole wall (pseudo-Rayleigh and Stoneley) .‘O These waves travel at different velocities, varying from about 25,000-4,000 ftísec.

A short time after the transmitter is pulsed, the near receiver senses the arrival of the various wave fronts. A little later they are sensed at the far receiver. The normal sequence is that shola n in Fig. 5-27, with the compressional wave arriving first and the shear wave next. These actually travel as a compressional wave in the mud from the transmitter to the borehole wall, a body wave (compressional or shear) in the formation, and a compressional wave in the mud from the wall to the receiver with the initial energy following the critically refracted (minimum time) path indicated in the figure.

Closely following the shear wave is the guided pseudo-Rayleigh wave, followed by the direct mud arrival and the guided Stoneley wave. The direct sonde wave is mixed with the mud and Stoneley waves. Every effort is made in the design of the sonde to make its transmission as weak and as slow as possible.

Of primary interest are the compressional and shear waves. By definition, a compressional wave is one in which the particles in the medium are vibrating in the same direction as the energy is propagating, in this case parallel to the borehole axis. A shear wave i s one in which the particles are vibrating at right angles to the propagatingdirection, in this caw perpendic- ular to the borehole axis. Compressional waves travei about 1.7 times faster than shear waves.

Standard Sonic logging tools at present measure only the compresional travel time. To effect this, the transmitter i s pulsed once and an electronic circuit measures the time elapsed to the first negative excursion of the compressional arrival at the near receiver (Fig. 5-27). The transmitter is pulsed again and the circuit measures elapsed time to the far receiver. The difference in arrival times is computed and divided by the span (in feet) between receivers, The result is presented on the log as formation transit time in microseconds per foot. Accuracy of the measurement is quite good, approximately f 0.25 pseclft.

Compressional travel times vary from 40 psecift in hard formations to 150 psec/ft in soft ones. Corresponding velocities, which are the inverse of

e sode an o zri nd


transit times, vary froni 25,000-6,600 ftisec. For comparison, the travel time in water (or drilling mud) is approximately 190 psecíft. A logging tool will occasionally read this value in an extremely large hole washout where

measure shear wave travel times also. This technique is evolving at present and is discussed under Long-Spacing Sonic logs.

THE BOREHOLE COMPENSATED LOG

The first-generation Sonic tool of the type illustrated in Fig. 5-27 suffered from the fact that when a hole enlargement or contraction was spanned by the receivers, the travel time across the mud was not the same at both receivers and therefore the recorded formation travel time was in error over a depth interval equal to the le span.”Very hashy logs were

1960s by the current Borehole Compensated Sonic* shown in Fig. 5-28,22 This tool consists of two transmitter-receiver arrays, one inverted relative to the other. A sequence of four time measurements is made as indicated. The lower transmitter is pulsed twice in succession, and the time interval (T2 - Ti) between arrivals at receivers R, and RI is measured. It will be abnormally long with the hole enlargement indicated. Then the tipper transmitter is pulsed twice, and the time interval (T, - T,) between arrivals at receivers R, and R, is nieasiired. It will be too short in the exan-iple. The correct travel time is obtained by averaging the two readingr. The averaging technique also eliminates incorrect reaclingc due to sonde tilt in the hole. \Vith this technique the quality of Sonic logs in rough holes is vastly improved.

The Schlumberger RHC tool has a spacing of 3 ft between transmitter and near receiver and a span of 2 ft between receivers. The transmitters are pulsed a total of 20 times per second so that five complete measurements are made each second. Logging speed i5 5,000 ftihr, which means a measurement is made about every 3 in. of hole. Normally the Sonic tool is run centered so the contributions to a receiver signal from different sides of the hole will be in phase (if the hole is round) and the signal-noise ratio will be maximized. The tool can be run off center, but significant degradation i n the signal-noise ratio must be tolerated.

‘Designated RHC Sonic b> Schliimberger, BHC .4coiistilog b! Dresser-Atlas, .4couctic V C ~ O C - it’ log by Welex. and BCS log by Gearhart


- - 3 - - -L 1 _ _ _ - 3’. _-I--- 1 - -2,’ 1 - __ I_

-

4 C$ LT R4

t Miid

Formatiori -X-

Measurements from lower transmitter LT

Output from receiver 1

Measurements from upper transmitter UT + I

Fig. 5-28 Borehole Compensated Sonic logging system (after Thonias, courtesy SPWLAj

Vertical Resolution and Depth of Penetration

The vertical resolution of a Sonic tool is well defined. It is the span between receivers, that is, 2 ft for the standard BHC tool. The depth of penetration is not well defined but is very small: it is controlled by the basic frequency in the waveform (= 25 kHz). For a homogeneous formation it is in the range of 1.0-2 in. and is independent of the spacing or span of the tools.


If, however, the forniatioii is not homogeneous but contains an altered zone of slower \docit!. next to the borehole, the depth of penetration of the ineasureriient can be increased by increasing the spacing, in the sense that the first arrival of energ’ at a rweiver can be made to come from the deeper unaltered zone. This is discussed under Long-Spacing Sonic.

Log Presentation

Typical presentation of the Sonic log, when run by itself, is shown in Fig. 5-29. The interval transit time, in microseconds per foot, is recorded across Tracks 2 and 3. Short transit times are to the right and long transit times are to the left, such that increase in porosity deflects the curve toward the depth track consistent with Density and Neutron recording.

In the depth track are small pips, representing integrated travel time of 1 rnsec between each pip. Larger pips are recorded at 10-msec intervals. These are useful in comparing Sonic logs with seismic sections. Usually a caliper curve is displayed in Track 1, obtained from the combination cali- pericentralizer run above the Sonic tool (another centralizer being riin below). When a Gamma Ray log is run simultaneously, it is also recorded in Track 1. If a resistivity tool is run simultaneously, as is often the case, the resistivity curves are displayed in Track 2 and the Sonic travel time is restricted to Track 3.

A good check on the accuracy of a Sonic log is to observe the reading in casing. It should be 57 psec/ft, the travel time of steel. The log may not jump immediately to this value on entering casing because there can be a drastic change in signal amplitude to which the system (or engineer) must adjust. The reading is most reliable in uncemented pipe where the casing-borne arrival will have good amplitude and will always arrive ahead of formatiori- borne signals, no matter how fast. The opposite can be true in cemented pipe.

Noise Spikes and Cycle Skipping

T\vo effect5 may cause erroneous spikes or shifts to occur on Sonic logs. The first is noise triggering, as illustrated by Fig. 5-30. Road noise is generated by tool centralizers rultbing against the borehole wall; a noise pulse may occasionally trigger a receiver circuit ahead of the desired arrival. This is most likely to occur at the far receiver since the trigger circuit must listen longer for a signal than at the near receiver and must also be capable of picking up a smaller signal. (Both detector channels are closed for the first

L 1

r; w 8 z - 5 u (o 3 3

I e R

1500 ." I

I

I I I I I ,

f

?

N

3


require input of elastic properties of the grain material, the pore fluid, and the empty rock frame.23 These are not known with certainty in anyparticu- lar case, especially the empty rock parameters. Therefore, it is necessary to resort to empirical relations between velocity, or transit time, and porosity.

The empirical relationship universally used until recently is the Wyllie time-average formula. It is based on laboratory observations and states. in effect, that the travel time of a compressional wave through a block of porous rock is the same as if all of the matrix material in the rock were pressed into a solid piece at one end of the block and all of the pore fluid were gathered in the remainder of the space with the compressional wave travcl- ing through the two portions in This leads to the simple expression

(5.15)

w

t Cp = porosity t, = travel time of the flu cupies the pores t,,, = travel time of the sol atrix

= travel time of the porous rock

Solving for porosity

9 = (t - LJ/k - t-,,, 1 (5.16)

Using this expression porosity can be obtained from the log-recorded trave¡ time. t. provided tf and t,,,, are known. The fluid in the zone of investigation i s typically mud filtrate. Consequently, tf is riormally taken as 189 psecift in fresh mud. In salt mud a value of 185 may be used. Matrix travel times vary from 40-50 pseclft, depending on lithology.

The solid lines of Fig. 5-33 provide a graphical solution to Eq. 5.16. Log- derived transit time is entered on the horizontal axis, a line is projected vertically to the appropriate matrix velocity, and porosity i s read on the vertical scale opposite the point of intersection.

As an example, the zone at 12,558-12,564 ft on Fig. 5-29 reads a travel time of 66 psecift. Fig. 5-33 gives a porosity as low as 8 5% if the matrix i s sandstone or as high as 18 % if the matrix i s dolomite. Clearly, the lithology must be known to obtain accurate porosity values. Even within a given


lithology there is a range of possible matrix travel times, as indicated at the bottom of Fig. 5-33. Local knowledge dictates the value to use - although the deeper the burial of a formation. the lower t , , is likely to be.

A comparison of Fig. 5-29 with Fig. 5-34, the Density-Neutrori log over

defining lithology as well as porosity. Using the crossplot chart of Fig. 5-21, porosity and lithology for selected intervals of the D-N log are listed in columns 2 and 3 of Table 5-3.

$e, cleart); 3hOWS

Y . (ft/sec) 1 <I (microsec/ft)

Sandstones 18 000-19,500 55 5.51 3

timertones 21 o00 23,000 47 6-43 5

Dolomites 23 000-26 O00 43 5 38 5

F ig 5-33 Determination of porosity from transit time b y time-average method (courtesy Schiumberger)


12500

i

I

1260ri

I -

Fig. 5-34 DensityNeutron Schlumberger)

Cuniperisalrd torrn,ilian aenslty porosity

35 10 15 0 l i ----T--- I

log. same interval us Fig . 5-29 (courtesy

,' '


TABLE 5-3 COMPARISON OF SONIC AND DENSITY-NEUTRON POROSITIES

Density- Range of Sonic Porosity with

interval, Ft Porosity, % Lithology Porosity, % Fig. 5-33 Fig. 5-36 Neutron Density-Neutron Sonic D-N LilhOlOgy

1 2,466.- 70 14 Limy doloriiite 6-14 12 14

12.525-28 16 Lirriestone 8-16 12 14

12,558-64 18 10-18 16 17 dolorriitic 12.62U-30 6 Limestone 3-1 1 6 5 6

12,480-52O O Anhydrita O-? 5 O O

Limestorle. slightly

In coliirrin 4 is the range of Sonic porosities p~issiible without knowing litliolog>., asaiirnirig t,,,, for sandstorit: is 10,50í) ft/sec at the depth indicated. The raiigt, is iriiolerably large. However, using the D-N IitholoL?., Conic porosities are ttiosii @\#en in coltiinri 5. These agree reasonabl- ~ ~ 4 1 with the 13-N \ ~ ; ~ l u i ~ s of cüliiiiiii 2 .

Correction for Lack of Compaction

The tiriit.-a\wage relation holcís cliiitt ' well in coiisolidated or \ve11- compacted fornintions. Typically, tliese tidve transit tiincs less than 100 psecift. Howww, serioiis errors arise if the relation is applied nithotit modification in sliallow, iinconsolidatcd sands." If the effective pressure on the rock f r m i e w o r k (ovc,rl>iirde?i--h-dr(istatic) is less than atmiit 4,000 psi, ivhicti is tlie case at drpths lcss thari at>ciiit 7,000 ft, tlie sand tias not reached its fully compacted rigidity and therefore its velocity. Travel tirnes in tiri<~ori-iI)a"ted sands iiial- reach as much as 150 psedft, which comwt to prosities far above tlici kriow,ri riiaxiriiuni of -$O ',?J .

To take care of this situation, the porosity conipiited from Eel. 5.16 is di\.ided b y a compaction correction factor R,,,, as indicated in Fig. 5-33. The factor variw from 1.0 to as high ab 1.8. It can be estimated by observing the transit times of shalc~;; tjh, adjacent to tlie zone of interest. These transit tiiiwc also increase writh lack of compactiori. Wtieri they exceed 100 psecift, E,:) is cipl)roxiiriatcd by

H, ,' = t,,,i 1 O 0

The dashed lines of Fig. 5-33 are appropriate to use for B,, values exceeding 1.0. B,, is never less thari unity.

Fig. 5-3<5, taketi from a Gulf of hlexico log, exemplifies the need for a conqxictioii correctiori. The water-bearing sand at 4,100 ft averages about

(5.17)

P

O o

P

N

O

O

P

o

O

O

P

P

O

O

P

vi o e

o,

O

O

P

4

O

0

Por

osity

, a'o W

D

C

=o

r

AN

O

0 I

I

154 ESSENTIALS OF. M O D E R N OPEN-HOLE LOG INTERPRETATION

range 1Oo-i50 [tsecift. porosity values obtained from i'ig 5-36 for. dit: selectcd intervals of the log of Fig. 5-29 are given i n colunin 6 of Table 5 - 3 . They agree very well with the Ilerisity-Neutron values in colunin 2.

Gas-Bearing Formations

The presence of gas in the pore space of a rock will increase the Sonic transit time over its value in the same liquid-saturated rock. Cas i s \ w y conipressible; when it replaces pore liquid, it lowers the rock rigidit!, more than its density and decreases Sonic velocity.

The decrease i n velocity is alrriost nil in the deeper low-porosity forniatioris where pore volume is l o ~ v and compaction pressure is high, ~ . l i i c h means that pore fluid contributes little to rock rigidity. However, it can be as high as 40 7" in shallow, high-porosity formations where pore voliinie is large and compaction pressure is minimum - in which case pore fluid tias a Iniich larger contribution to formation rigidity.

Fig. 5-37 shows calculated velocities for sands of porosities 30 %' to 26 :'u at depths from 2,000-10,000 it as li\~drocarbon saturation increases from O to 1.0 (proceeding froni right to left in the horizontal scale)."',,'" Cas satiiration is represented by the solid lines, and oil saturation is represented by tlie broken lines. As gas saturation increases froni O to about 15%, the velocity drops drastically, wliich is a result of the fluid cuinpressibility rapidly increasing (Fig. 5--38). IVith further increase in gas saturation, however, velocity gracluall!~ increases again since the rapid decrease in density of tlie pore fluid inore than cornpensates for the additional increase in fluid cwn- pressibilit).. The niaxinium velocity decrease predicted is about 1.0 ';ó for the 10,000-ft case arid 35 'L for the 2,000-ft case.

IVhether the Conic will sense thc presence of gas depends on how iiiucli

residual gas is left after invasion in the 1 in. or so of formation 1)eiiig investigated by the tool. In medium- to high-porosity gas-bearing forrna- tioris, one would expect a residual gas saturation of at least 15"L i n t - 2 ~ flushed zone so that gas should be seiised by the tool. This is implied i n the gas-saturated sandstone line of Fig. 5-36 where tlic shift in travel tinit, from liquid tu gas is consistent with tlie predictions u! Fig. 5-37,

Fig. 5-35 skiow an example of an unusually large gas response. 'The upper 80 ft of the sand from 4,520-4,715 ft is gas bearing with approxiniately 90 '% gas saturation in the nonirivaded zone. The Sonic log s h ~ \ v ? : a n increase of abut 50 psecift in travel time, or about 35 % as predicted by Fig. 5-36 (extrapolated). IIowever, signal amplitudes i n uncompacted sands can become extremely low sí) that increased travel tirnes can be due to cvcle

a

I

/' /

/ I -0- LA--- - I 10,OOO FT I

/ * 0

. -

- GAS SAND -- O I L SAND

' O O. 5 ! .O BRINE SATURATION i S I

W

Fig. 5-37 Calculated sand velocities with varying oil and gas saturations (after Domenico, SEG)

d.ipping 01, in the eutremc-, to the tool rceding niud travel time. The lattri íippeiirc to be the case in the 4,535- 4,550-ft and 4,572-4,580-ft intervals of the example.

Ovrrail, the Sonic log cannot be considered a reliable gas indicator, either because the Lone of investigation can be completely íiushed of gas or because the more usual increase of 5-1O'Ro in travel time appears siniply as


-1 U

1 a.10 0.1 0.3 9.5 0.7 0.8 0.9 1.0

Fig 5-38 Compressibility of a gas-water mixture as a function of water saturation (after Domenico, CEG)

an increase in porosity of similar magnitude. In the case of oil, Fig. 5-31 clearly shows that a residual oii saturation of 30% in the flushed zone will have much less effect than gas on the Sonic log and therefore will be even less distinguishable.

Secondary Porosity

The need to have a Density-Neutron log that can give both lithology and porosity in Iow-porosity regions does not completely negate the value of a Sonic log. It is useful as an indicator of secondary porosity, which occurs primarily in carbonates.

The Sonic log tends to ignore porosity in the form of isolated vugs or channels. Loosely speaking, the tool will not see such porosity if a vertical

157 POROSITY LOGS *

slice of the circumference of a hole between the two receivers is free of such vugs or channels: the first arrivals will follow that path. On the other hand, the Density and the Neutron, which respond to average porosity in their +in. to 10-in. depths of investigation, will average vugular porosity with int

between the D-N and Sonic porosities (& - &). Asecondary porosity index may be defined as

Secondary porosity may therefore be

SPI = (+‘h - #J/+d” (5.19)

The greater the SPI, the higher the permeability is likely to be for a given total porosity, assuming the vugs or channels are interconnected. This is valuable information for estimating formation productivity.

Porosity determinations, however, must be very accurate if the SPI is to

Hole Enlargement and Formation Alteration Effects

Sonic logs are very tolerant of borehole size variations within the norma¡ 6-12 in. range. However, serious errors can occur in transit times recorded with a standard BHC-type Sonic tool in larger holes or those with appreciable formation alteration.

Considering Fig. 5-27, one can readily visualize that if the hole diameter is too large and the Sonic tool is centralized, the mud wave proceeding directly down the hole will arrive at the receiver ahead of the formation compressional wave. The recorde¿ travel time will then not reflect formation travel time.

The maximum hole size tolerable depends on the shortest transmitter- receiver spacing and the formation travel times3’ For the BHC tool with a near receiver spacing of 3 ft, the maximum tolerable hole size is 14 in. for a 150-psecift formation and 22 in. for a 100-psecift formation. Enlargements of this nature do occur, particularly in shallow uncompacted formations.

Formation alteration is similar in its effect. Typically, such alteration is the swelling of water-sensitive shales in contact with fresh drilling muds.32


The s\\~elliiig begins at the 1)oreliole \\,all and slowly progresses oii t\vard. The s\\dleri shale in effect has higher porosity anti therefore lower velocity than t l ie iindistiiit>ed shale l~eliinci i t . I f the altered zone has appreciatile thickness a t the tiriir of logging. the first a r r i \d of e i i e r ~ at the receivcrs can be tlirough that l a y r arid the Sonic tool \ id1 rrüd an at>iiorrrially long transit time. A striking example is sho\vn in Fig. 5--3c3 u h w c there is a progresyivc incrrase inshale travel tinies (from approxiniatel!. 125 to l-tOpsecift at 9,:3%) ft, for exaniple) over an interval of 3 to 35 days after drilling. Little alteration occiirs in the sands, such as at '3,292 f t . ?'his is fairl!. typical, tlioiigli shallo\v sarids soinetiines exhibit alteration.

l'he niasimiim alteration depth t<.)lerablt~ depeiids on borcthole size, near recei\.er spacing, aiid hot ti altered aiid unaltered zone trkiirel tiii1c.s. The staridard BIIC tool ii i 10-in. hc~le can c~opc ~ . i t h only 2--3 in. of alteration in forniations of 100-150-psecift tra\>rl tinies.

The solution for Iioth hole d:irgeriivnt and shale s\vellingprobleiiis i j to utilize a Sonic tool with lorixer transriiitter-rccei\.t.r spacin S- f t niininium spacing can tolerate nlniost an" coiiceivable iiole c3iarnett.r and can cope Lvith 8-13 in. altrratiiin depths under the conditions cittd pre\.ioiisl>,.

In past !rears the main application of long-spaced tools has hem for geophysical piirlmses. \\.herein Conic logs are used to depth-calibrate seismic sect ims through the niediuiii of s>.iithctic scsisniograiiis. Travt.1 tiiiie anoiiia- lies <:aused 11y shalr alteration x i d hole enlargement (the t u o often go togctlier) can cause false reflection, {JII tht: synthetic seismograms and poor time inatch \<.ith the seismic section. I n sha l lo~ . soft rock areas such as the U.S. Ciilf Chis t Lvliere shale alteration is cornriion. a long-spacing tool stioiild :il\va!x tie run rat1it.r than the standard BI-IC.

Receiitly, siiccess has 11et.n achie\.cd in shear \va\.e logging and in logging throiigh casing ivith loiig-hpacfd roJidex. Increased tise of siich tool:; for thear purposes i n the future can I)<. t'xpecteci.

I

THE LONG-SPACING SONIC LOG'

Fig 5 IO \Iron 5 \clic.iiiaticall> the arr,in-;eincwt of t l i e Cchliiiribtqt.r I X tool, Tu o transmitters Are sped 2 ft apart at the bottom of t h e tool arid two receners are $paced 2 ft apart dt the top u ith 8 f t of space between t l ie nearest tranmitter m d receiver. Ll'ith thi5 arrangement two long-spacing

POROSITY LOGS 159

logs are recorded simultaneousl!., one wi th 8-10 ft spacings arid one with 10 -12 ft spacings.

Borehole compensation is accomplished by depth rriein«rizatio~i rather than b y the inverted array technique. To illustrate, the tra\'el time at depth

-JXA 9300

150 A t , )I secft

1 O0 50

9350

Ruri #5 _ _ -__ Run #6 . . . . . . . . .

-- Fig. 5-39 Effect of shale alterution; 'days after drilling (after Blakerr\an. courtesy SPWLA)

3' 13

25

35


R:

I a f t

I

'I i I I I I I I 1

I 1-2- 0

Fig, 5-40 Schlumberger Long-Spacing Sonic sonde

POROSITY LOOS 161

level A is first measured with the two receivers spanning that level. Trans- mitter T, is pulsed twice in succession and the time difference (T,R, - TI&) corresponding to compressional arrivals is placed in memory. If the hole diameter is different a t the two receivers, themeasured time will be in error.

only receiver R,, the time difference (T,R, - T,R2) i s measured; it will be in error by a like amount but in the opposite direction. This time difference is then averaged with the previous value, retrieved from memory, to provide an 8-10 ft spacing travel time compensated for borehole variation.

Using transmitter T2 in the first position instead of Ti arid receiver R, in the second position instead of R2 provides a 10-12 ft compensated log in like fashion.

the BHC curve reads a transit time too long because of shale alteration. In the sections above and below the sand, it reads much too long became of hole washout. By contrast, th S log reads correctly.

Anomalous Triggering Effects

Unfortunately, longer spacings lower signal amplitudes at the receivers and aggravate noise triggering and cycle skipping. When this happens, spikes occur in pairs, spaced 0 2 / 3 ft apart, because each T-R time measurement is used twice for time computation: once at the level it is taken and again 925 ft higher in the hole.

This behavior is illustrated in Fig. 5-42. The 8-10-ft log is laheled DT and the 10-12-ft log is labeled DTL. An incorrect reading recorded on the S- f t spacing will cause two spikes of like polarity spaced g2/3 ft apart to appear on the 8-10-ft log (pair D). Similarly, an error on the 12-ft spacing will cause a similar pair of spikes in the 10-12-ft log. An error on one of the 10-ft nieasurements, however, will cause a spike of one polarity on the 8-10-ft log and a displaced spike of opposite polarity in the 10-12-ft log (pairs A, 8, and C ) . These anomalies can be smoothed out by eyeball averaging or eliminated more precisely by computer pro~essing.~'

We can also expect to see improper borehole compensation when sonde sticking causes erratic downhole motion. If sticking occurs, as evidenced by

I 1.

-

r-- --

I I

I I

1 i I

1 1

13800 13900

-.-

'I

. _.

-1

I cn o

m r (?

Q

/I

I

Q

I


recorded travel time will be in error if the hole configuration is different at the two levels. No easy solution to this problem exists.

The accuracy of LSS measurements can be checked. as with thestandard

signal amplitudes in ccmcnteci caiiiig are too m a i l at the !ong spacings i i s v d , Derivation of porositc from LSS travel times is exad:: tlie same as for the

standard BHC Figs. 5 - 33 and 5-36 are applicable.

Shear Travel Time Measurement

The desirabilit! of measirriiig shear wa\e velocity in addition to compressional velocit) has long been recognized. The basic reawn ir that. \ k i t h

the two velocities. the, fundamental elastic constants of the medium can be h as sand- plirationc

Shear wavc ineasurement requires data ;icqnisit inn u ith long sondes, digitizing of sonic wa\e form4 a t the ~vvelísitc, and sophi wave form processing

The longer the spa r the Eeparation i n time between thP compresional and ahear xv:avf= packets and thc easier it i s to sensc acciiratel! the later shear ariivals. Fig. 5-43 illustrates the point. It i s a variable- density recording of rweiver wave forms ai 8-ft spacing. Once per foot of hole. the receil er sipnal modulates the intensity of an electron bcani sweep- rng across the face oí 'I cai-hodc I a) tube !CRT) M i th the 511 cep <tarking v, hen the trammittcr i \ pulwd. The modrilated s\\ peps recorded on a n i o v i n ~ film to pro\ ide a díspla!r in which wa\.e anipliti~des are tiandated into black-and-Xvhite contrasts.

The compressional arrivals are seen clearly at about O 3 zncec and the shear arrivals at approximately 0.0 msec. They are well separatd, which facilitates shear wave extraction.

While shear transit time can be determined from variable-density displays, it is a tedious and imprecise process. Accurate determination requires that the wave forms be digitized a i 5- to IO-psec intervals and processed through correlation and filtering techniques to suppress the direct compressional wave and reflections associated with it and to enhance the shear arrival.36 A tremendous amount of data must be handled, so the processing must be done in computer centers at present.

POROSITY LOGS 165

-- h"iLL'CECONDS c P - u - '

k g 5-43 Variable derisitv display of waveform? with 8-A tfancmitter receiver spacing (after Siegfried and Castagno, courresy SPWLk)

Schlumberger's computation method, direct phuse det erm i t iu t ion . is com- mercially a~a i l ab ie .~ ' It computes compressional and shear travel times, their ratio, and energies of the compressional and shear packets.

Fig. 5-44 i s an example of compressional and shear logs recorded in a sand-shale series.38 Travel time curves are displayed in Track 5, and the ratio of shear to compressional travel times (denoted AT,/ ATc) is plotted as a solid line in Track 4.*In Tracks 1, 2. and 3 are the grain densities, fluid

'Along with the derived Poisson's ratio (dashed line).

h

a

8

C

4 68 ESSENTIALS OF M O D E R N OPEN-HOLE LOG INTERPRETATION

ca ,

tant not to exceed the fracture pressure or circulatiori will be lost. On the other hand, when fracturing tight formations to improve their produetivi- ties. fracture pressures must be exceeded. Fracture pressurt. may be \vrit- ten'"

I

POROSITY LOGS 169

(3.25)

where

Po = overburden pressure (0.9 to 1.0 psiíft) P, = pore pressure (measurable with a formation tester) a = constant, approximately 0.5

The fracture pressure is very dependent on Poisson's ratio. Conse- quently, it i s important to have an accurately measured value rather than a

ly obtained as indicated above, Fig. 5-44 shows that Pois- vary froin 0.15-0.33, a wider range than the empirical

rnethod predicts. Improved determinations of other derived parameters. such as fracture

widths and heights and rock drillability. can be expected with more precise knowledge of the elastic moduli."

Lithology Identification i

The ratio of shear to compressional travel time appears to reflect lithology for the three major sedimentary series. Typical values for clean formations are

Lithology

Sandstone Dolomite Limestone

f 4, 1.58-1.78

1.8 1.9

as the P, curve of the Litho-Density tool whose values increase from 1.8 to 5.0, progressing from sand to dolomite to limestone. On the other hand the tJt, ratio can be useful in wells drilled with barite-loaded mud where the Fe curve is not usable.

The constancy of t,/t, implies that shear travel time has much the same dependency on porosity as compressional velocity. In fact, the abscissa scale of Fig. 5-36 can be multiplied by the appropriate ratio to produce an approximate porosity vs shear travel time curve for each lithology.

Clay Indication

In sand-shale seque es the presence of clay and silt increases the t, it, ratio from a nominal value of 1.6 to as high as 1.9, as illustrated by Fig. 5-44. Once again th of variation is not nearly as great as for the

nce that shear attenuation may as to the content. If this could

distinguish between dispersed and laminated clay, it wouid significantly enhance formation evaluation.

Gas Effect

The presence of gas i n pore space should decrease the shear wave travel time. This is because shear velocity, V,, of an elastic solid is given by

v, = (p/p)'h (5.26)

where p is the shear modulus and p is the density. Replacing liquid by gas in the pores should not affect the shear modulus, since fluids do not support shear, but it will lower the density. A gas saturation of 15 % in the flushed zone should lower the travel time by about 5 % . It will increase thecompres- sional trave1 time by about 10 % , so the t,/t, ratio should decrease by about 15%. Fie. 5-45 shows a 10% decrease in ratio, from 1.8 to 1.62, most of

c.7

A statistical plot from several wells is shown in Fig. 5-45. The limestone and dolomite ratios are independent of porosity over the 0-20% range

which was due to change in t,. The conclusion is that the shear travel time log will be relatively insensitive to gas.


. Limestone

SANDSTONE ’

G A S

-L. 100 1 I

160 140 DiS 80 1 O 0 120

Fig. 5-45 Comparison between lithology and shear to compressional travel time ratio with data from several wells (courtesy Schluniberger and SPWLA)

ELECTROMAGNETIC PROPAGATION-MICROLOG COMBINATION

?’lit, 1; lec t ro ni agne t i c Propagation- hi icrolog (E 1’T - M L) coni hi 11 at i oi 1 is a marriage of one of the earliest and siniplest t:lectrical tools (ML) nitli one of the latest and most sophisticated electromagrietic tools (EP?’). What they have in common is that

both are pad devices both have extremely high vertical resolution and very sliallow depth of investigation, a few inches in each case both display information concerning hydrocarbon producibility both are best suited to fresh mud conditions logging speed for both i s approximately 2,000 ftihr


Use of Micrologs

The prime use of the .\licrolog is t o indicate pernieatdr Z C J I I ~ S . It does tliis iri great detail 11). sensing the presrrice or almrice of niud cake. I’rrriieable zones al lo^, invasion of mud filtrate, i+,liich results i n rniid cake t)iiildup; iiiiperriic~tt>le zones do r i o t invade. ?‘lie LIicrolog is still the best perineable zone or “ s a d coiint” indicator toda!,. I ~ I C J \ W W ~ , it gives no indication of the i n a gn i t u de of t 11 e per ni i ~ h il i t !. .

Use of EPT Logs The Electromagnetic I’ropagatioii Log essentially gives the water-filled

porosity in t l ie first few inches of forrriation, largely ignoring tiydrocarbon- filled porosity. * Its principal use is for determining the water saturation of the fliislied zone. ‘Iliis is done b), comparing the EPT porosity, dFt,; with the total fliiid porosit!., 4, ohtnined froin the Density Neutron coriihiriati«n. Flushetl-zone water saturation is approximated b!.

S,<) = di;,>/@ ( 5 .27)

This üpproacli is indcpmtlent uf thc saliriity of the nrater in the flush~d zone. It does not reqiiirr. a kiioudedgc~ of mud filtrate resistivit!, or of \vhether ~ i i u d filtrate curripletely replaced connate water in the flushed Y.011e.

Applications of the EP‘T fall into the following categories. 1. lhtcrriiiriation of hydrocarbon movability by comparing the

fliishecl- zone saturation, C,,,, with tlie iindisturbed-zone saturation, S,, . Xlo\.ed hydrocarlion, as a fraction of the pore space, is (S,,, - S,,). The teclini(lur is most applicable to fresh-~niicl situations where El,,, logs ( M l L , PL,, or LISFL) are seldom rim and where, if run, may not provide reliable

2 . Detection of li!.drocarbons in fresti\vater areas where electrical logs caiinot distinguish water from oil. ?’his situation is ~on1111011 to stiallon prodiicirig areas, often Lvliere oil is fairl!. heair!,. The liea\+r the oil the greater the residual oil saturation in the fluslied zone, so the more readil!.it is detected as a difference betiveen &,, and d. Heavy oil regions of California a r i d the Athabacca tar sands are exaniples. In siich areas there is SO little displacement of oil b y niud filtrate that S<<> approximates the actual water saturation, S,, , of the resenwir.

3. Delineation o f 100 $1, water-hearing zones, \vhere ¿ # J ~ ~ ~ = 4, xvliicli then can be iised for determination of R,, . Wet zones can readily he picked

S,,,\~alilcs,

‘.4vailahIe only from Schlurnberger at preseiit

4 72 ESSENTIALS OF MODERN OPEN-HOLE LOG INTERPRETATION

on resistivity logs in porous Miocene sands, for example, but may not be evident in carbonate sequences where porosities are lower or in regions where uplifting or overthrusting of sediments cause R,, values to change rapidly.

THE EPT-ML SENSOR ARRAY Fig. 5-46 shows the portion of the EPT-MI, tool that contains the sen-

sors. On theupper part is a fixed pad, about 1-ft long, with EPT transmitters and receivers. This pad is forced to ride the side of the hole by a strong backup arm on which a Microlog pad is affixed. The arm also measures gross hole diameter. At the same level but in line with the EPT sensors i s a small

TRANSMITTING ANTENNAS

SMALL ARM CALIPER *y

t JJJ r*- -

LARGE ARM CALIPER

MICROLOG PAD

Fig. 5-46 Electromagnetic Propagation-Microlog sonde (courtesy Schiumberger, O SPE-AIME)

POROSITY LOGS 173

weak arm that measures the rugosity of the wall traversed by the EPT pad. The sum of the large and small arm displacements is recorded as hole diameter. Both arms are hydraulically retracted for descent into the hole

The EPT-ML tool can be run in combination with the GH-CNL-FDC (or LDT) array, as shown in Fig. 5-47. Combined tool length is approximately 80 ft: logging speed is 1,800 ftihr. A high-speed telemetry cartridge transmits all information to the surface in digital form where it is processed in the logging unit computer. This telemetry slsteni. bccoming universal in Schlumberger, greatly extends tool combiriabilit! .

THE MICROLO

can conform to hole irregiilarities. It is important to the quality of the log that pad contact with the wall be sufficiently good to prevent drilling inud from directly contacting the electrodes.

A constant survey current is emitted from the lower button. This current passes into the formation and returns to the sonde body. A Micronormal resistivity curve. denoted R2 or MNOR, i s recorded by measuring the poten- tini betneen the upper button and sonde mass. Similarly, a llicroinverse cunx,. clerioted 11, ~ , or i\IINV, is recorded by measuring the potential hrtween the middle and upper buttons. The depth of irivestigatíori of the hlicronormal is approximately 4 in . and that of the Xlicroin\er\e 1.5 in.

When no mud cake is present. as in imperrtieable zones, both ('tirves read the same \ aliie. Thus. the curves overlay in $hales or in impermeable sand< or carbonates if resistivitj, i i not too high. \'l'hen mud cake exists, ho\vever, the cake generally has lower resistivity (essentially higher porosity) than the flushed formation immediately behind it. Since the Microinverse curve has shallower penetration, it is more influenced by the niud cake and reads lower resistivity than the Micronormal. Positive separation betwecn the curves, with MNOR > hfINV, indicates permeability.

"Denoted Microlog (ML) by Schlurnberger, Minilog by Dresqer Atlas, Contact log h) R'elex, and Microelectric Log (MEL) b) Ctlarhart.

174 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETA'TION

Example Log Fig. 5-4s is an example of i i Slicrolog riin iii coniI)iriati«n rvitli G R

Demit'. and Neutron logs. Iri Track 1 is recorderi the c;iliper arid CR. Tlie latter slio.r\,s t ~ k . 0 niajor sands ;it inter\,als A and 17 a lo~lg \viti1 indications »f thin sands at B, C, and 11. The Slicrulog <'iir\ys are recorded in Trüc.k 2 , Positive separation, indicatrd I)? tlie dot coding, clearl!. sliows tlie prrine- alde intervals, nariiely the top 7 ft of A, ahout 2 ft in B, aIid 23 ft iii E , for a total sand count of 33 ft. Zone C is indicated a s tiglit iind zoiie D as irnperrrie-

______ -- -___I

T E L E M E T R Y 8

GAMMA RAY

ELECTROMAGNETIC

PROPAGATION

~~

POWERED CALIPER MICROLOG

COMPENSATED NEUTRON

FORMATION DENSITY

-GR MEASURE PT.

I - - E P T MEASURE PT.

--- M L CALIPER MEASURE PT.

- - - -CNL MEASURE PT.

+-FDC MEASURE F T

- --- Fig. 5-47 EFT-Microlog in combination with FDC, CNL. and GR (courtesy Schl urn berger)


4 76 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION

able. Readings greater than 10 times the mud cake resistivity, R,, (estimated as about 1 ohm-m in this case), indicate tight zones. Separation of the curves at high resistivities is meaningless due to current leakage around the pad.

Mu An approximate value of mud-cake thickness can be obtained from the

chart shown in Fig. 5-49. Mud-cake resistib ity, R,,,, , must be obtained from the log heading and converted to the temperature of interest. Tlie ratios R2 /Rmc and R, iv/R1,, give a point of intersection frorri which niiid cake thickness, h,,,, can he estimated For zone E of Fig. 5-48, R," = 2.9 and R, = 2.1ohm-m. .issumingR,,,, = 1, h,,~isestimateclat % -92 irl. asiridi- cated on Fig. 5-49. I louwer . if R,,, were 0.5 oliin-111. h,,, woulrl he estimated as about iii. Consequrntl>. the procedure i s not very accurate

Fig. 5-49 Microlog interpretation chart (courtesy Schlumberger)

POROSITY LOGS 177

because R,, values on the log heading are not measured values but are derived from the Rmf measurement, assuming average mud properties. The main value of the Microlog is indicating the presence of mud cake.

d It isnot suited tosalt mud because the high -crondu

ity of salt mud tends to short-circuit the survey current. In early days before porosity logs were available, the Microlog was used

to estimate porosity."'R,, \ \odd be estimated from Fig. 5-49, S,, would be guessed and r$ would be derived from Eq. 2-12. However this was - and is - a very inaccurate method. The Microlog IS not a porosity device nor even an H,, measuring device. Microlog curves bear little resemblance to true R,, curves obtained with focused MLL, PL, or MSFL electrode systems.

THE ELECTROMAGNETIC PROPAGATION LOG

The EPT tool measures the travel tin waves propagating along the borehole in At the \.cry high frequency utilized. 1.1 x trabe1 of such waves is determined almost entirely by the dielectric properties of the formation and very little by its resisti\7ity.''z4' (The opposite is true for Induction and Laterolog tools, which operate at low frequ turn the dielectric permitivity is largely a function of the water content of the formation, '' "

Fig. 5-50 shows the EPT sensor arrangement. Two microwave transmitters (T,, T,) and two receivers (R,, R,) are mounted on a brass pad forced against the borehole wall. Spacing between transmitter and nearest reservoir is 8 cm and between the t x -o receivers i s 3 cm. The two transmitters are alternately pulsed, arid upgoing and downgoing travel times measured between the two receivers are averaged. This eliminates first-order effects of uneven niud-cake thickness, pad tilt. and instrumentation imbalances. Travel time is measured by sensing the phase difference in received signals at the two receivers. A complete measurement of travel time and signal attenuation i s made every l/mof a second and transmitted to the surface. There the measurements are averaged over 2-in. or 6-in. depth interval^."^^"

Vertical Resolution, Depth of Penetration, and Borehole Effects

Vertical resolution of the EPT log i s extremely good. It i s essentially the span between receivers, about 2 in. Depth of penetration is quite small, varying from about 1 in. in low-resistivity formations to about 6 in. in high-

o. h

- co h

.c

>

t

o $ 8 Y

Y L

" o 3 I z ,

O

z w


time of 11 nsím. Sand B is less uniform and reads a higher travel time. The remainder of the section is shown by the EPT as a series of intermixed shale and shaly sand stringers. The shale readings, those at highest travel times, average about 22 nsim.

decibelsimeter. Attenuati le that in shales is high, approximately 7.50 dbim. About 60 dblm of the attenuation is due to geometrical spreading of the propagating wave from its source. The remainder is primarily a function of the clay content of the formation. The higher the clay content, the greater the signal attenuation. The range is tremendous, a factor of 4,000 in this case between sand and shale levels for the 12-cm spacing. It requires that transmitter power be continuoilsly varied to maintain receiver signals in a workable range.

Also shown in Track 1 is the small arm caliper reading on a 0-10-in.

to 1,

is referred elsewhere for details.52 Travel time of microwaves in clean, lossless (low attenuation) porous

media is accurately given by the sum of the travel times through the coristit- uent portions. That is

t,, = 4 . t,f + (1 - $J) t:,r#, (5.28)

l&5 4J = &!>o - tPrnV($f - tl’”’)

Solving for porosity

(5.29) P

where

t,, = loss-free travel time of the medium, nsim

t,, = measured travel time of the medium ns/m A = measured attenuation of the medium, dbím t,, = travel time of the rock matrix, nsim t,, = travel time of the pore fluid, n d m

= [t,1’ - (A - 60)2i3.600]” (5.30)

POROSITY L O G S 4 8 4

Values of t,, and tPf for various rock matrices and pore fluids of interest are given in Table 5-4. Note that water has a much higher value than any other constituent. This is a result of its high dielectric constant. Values of the constant, relative to air = 1, are listed in the t, column. Propagation time in I1 n

t, = Jill, (5.31)

TABtE 5-4 RELATIVE DIELECTRIC CONSTANTS AND PROPAGATION i IMES FOR VARIOUS MINERALS

Mineral €1 1,. nslm ip,, nsim - Sards tone 4 65 7 2

Dolomite 6 8 8 7 Lmectone 7 5 9 1 Anhydrite 6 35 8 4 Dry colloids 5 76 8 0

4 16 6 8

- - - -

7 9-8 4 -

7 5-16 6 - Halite 5 6-6 35 -

5-25 Oil 2 2 - 4 9 Gas 3 3 - 6 0

its water content.

assumption of water-filled pore space. That is EPT porosity. (PEP, i s the porosity derived from Eq. 5.29 on the

\\..here t,, is the lossless travel time of water. This value is not constant but depends on temperature arid slightly on pressure. Over the range 100-300°F lvith usual pressure variation, it is given to sufficient accuracy by

(5.33) t,, = 31.1 - 0.029T

where T = formation temperature, O F .

travel timet, is 11.1 ns/m and attenuation A is 150 dbím. By Eq. 5.30 As an example we can compute &, for sand A of Fig, 5-51. Average

t, = [11.1’ - (150 - 60)213,600]’“ = 11.0 ns/m

This shows the correction for attenuation is negligible. The formation is known to be sand, so t,, = 7.2 nsim. Temperature at the 9,060-ft depth is estimated to be l iO”F, so that by Eq. 5.33


t,, = 31.1 - 0.029 x 170 = 26.1 risim

Application of Eq. 5.32 gives

r#JFP = (I1 . O - 7.2)/(%. 1-7.2) 0.20

Consequeiitly, the water-filled porosity of sand A is close to 20 . The actual porosity will be larger if the sand also contains hydrocarbons.

vs tlir> for sandstone and limestone, based on Eqs. 5.32 and 5.33. For clean formations of such types, the attenuation is negligible so that t,,, is the actual t,, value read frorn the log. It is apparent from the plot that matrix type must be known with certainty to obtain accurate porosity values and, in addition, temperature must be known to within 10°F.

Fig. 5-52 is a plot of

0.40

0.30

0.2c

T (PEP

0.1C

O 12 13 0 9 10 11

t, (nshn) - ~~~~

Fig. 5-52 Conversion of EPT travel times to water-filled porosity


Where the matrix is a mixture of several minerals, the matrix travel time is a linear combination of the matrix travel tinies of the constituents. That is

where V, is the fractional volume of the it,, component and t,,,, is the matrix travel time of that component. Relative volumes of usual components- silica, calcite, and dolomite-are best determined by the (p,,,), - (U,,,), method described in chapter 6, utilizing Litho-Density and Neutron logs. It is almost mandatory to run the LDT-CNL in conjunction with EPT for good interpretation.

EPT Flushed-Zone Water Saturation

The EPT porosity, &y, is the true porosity if all pores are water filled. It will approxiniate the water-filled porosity in formations containing hydrocarbons because hydrocarbon appears much like rock matrix to microwaves; travel times in the two media are similar. Consequently, an apparent water saturation for the flushed zone is simply

( U d = Qi,P/r#J (5.35)

where cb is total fluid porosity, generally obtained from Density-Neutron

The apparent saturation, (S,,,),, is clow to the truesaturation, S,,,, at high values h i t not at low values. To calculate the difference, Eq. 5.28 car: be expanded tor the case of both hydrocarbon and water in the pores to

logs.

tp,, = Sx,, ' 4 ' t p v + (1 - s,,,) ' r#J . tph + (l - r#J) ' t p r n (5.36)

where t,,,, i.s the travel time of the hydrocarbon and other terms are as already defined. This equation can bt, solved for S,,, and rearranged to give

C," = K + (1 - E;) (S,Ja (5.37)

where K IS ii coiistarit, depending only on matrix and fluid travel times, given by

'E; = (tpn - tph)/(tp - t i d (5.38)

from Eqs. 5.37 and 5.38. The difference is cniall at high valiies of (SAU), h i t appreciable at low values. In normal

Fig. 5-53 is a plot of S,, vs


hydrocarbon-bearing formations is greater than 0.7, in which case it is close to being correct. In heavy-oil situations, however, (Sx0)* could calculate less than 0.4. In this case the true S,, would be significantly higher than

Example of Moved Oil Estimation Fig. 5-54 is a recording of the basic Dual Induction-SFL, SP, GR,

Density, and Neutron curves over the same section of hole as depicted in Fig. 5-51. It is fairly obvious from the R,,, curve that sand A is hydrocarbon bearing and sand B is water bearing.

This is verified by the computed R,, curve in Track 1. From the reading of that curve in sand B, R, i s estimated to be 0.08 ohm-m. R,, in zone A

0.1 0.2 0.3 0.4 0.5 0.6 0.7 (%)a - - 0.8 0.9 1

Fig. 5-53 Conversion of apparent flushed-zone water saturation (S&, to true saturation S,,


averages 0.87, so a quick estimate of water saturation in sand A (see chapter 9) is

s,, = vRJR,, = Joo810.81 = 0.30

The if the sand is clean. However, if it is slightly shaly, as is probably the case, correction for shale would result in Neutron-Density crossover in the upper part of the sand, indicating the presence of gas.

r-

.--4u!m-alnllL- 5 T L U . I M W ---- EllUU_-> -___ o . COO@ ea 1 1 0 0 0 . m e . - - - - - 1 w- (PMWL - - - - I ñ.PO0Q eo b.booo 0.0

_ _ _ _ ~ m c 3 L . - - -- DtHX< > 0.0 I . ü@Q4

-@.*o a . 1800 I.;;&. . . .w. ..IWl.?. . . . . . . . . .. _ _ . . . .. ~~n..crrnnm _______ . ........ nrn.~c ..... > _ _ _ _ _ _ _ _ _ o. t o o 0 C O k. bOOQ 0.0

Fig 5-54 Resisttvity and Density-Neutron logs over ?he sume 1nte:val as Fig 5-51 (courtesy Schlumberger)

186 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION POROSITY L O G S 187

k'ig. 5-55 shokvs the water-filled porosity, ovrrlairi M.ith the total fluid-filled purosit)? + ohtniried by averaging tlie Density and Ntutroii porosities. The former reads 0.20 i n sand A, i n agreemint ivitli the previous cdculatiori, and tlie latter averages 0.27. The apparent flushed-zoma \vater saturation is then

(S,,,), =:: 0.20iO.2; = 0.74

The corrected S,,, valut. from Fig, 5 5 3 is 0.77. Moved hydrocarbon, as ii

fraction of the pore space, is

- A

9 '22

Fig. 5-55 Comparison of EPT porosity QEP with Density-Neutron porosity CP fo r same interval CIS Fig . 5-57 (cour?esy Schlun?berger)

S,, - C,, = 0.77 - 0.30 = 0.37

Sand A should therefore be a good hydrocarbon producer. On the other hand for sand 13,

Positiveseparatiori of the hlicrologciirvesinsand A, asshown iriTrack 1, clearly indicates permeability in that zone, corifirrning its potential as a hydrocarbon producer. Sand B is also permeable. However, all other thin shaly sands over the recorded interval are indicated as impermeable.

The ripper 4 ft of sand A was perforated arid flo\ved 70 klcfd and 30 bqid, confirming tlie log analysis.

= +, which verifies i t is water bearing.

Example of Detection of Heavy Oil in Fresh Water Areas53

Fig. 5-56 is a cornposite log of a shallow well in Kern County, Califor- nia, through a number of sands from which 12"-15" API gravity oil is produced. Forrnatiori water resistivities ín the area vary widely, froin 0.24 to 32 ohm-m at 75°F' (25,000 to 150 ppm), so that deterniination of water saturation from the electrical logs is virtually irnpossible. The SP in Track 1 is of little help in reflecting R,v changes or even in delineating sand-shale boundaries.

Porosities of various sands, labeled A to E, are fairly iiniforrn, averaging 38%) as shown i n Track 3. There is good agreement between Density- Neutron total fluid porosity + and porosity +,.,,,, from conventional cores. Track 3 also records &p. It averages much less than +, indicating considerable hydrocarbons in place.

values computed from the &,/+ ratio arid corrected as indicated by Fig. 5-53 are recorded in Track 2 along with residiial oil saturation S,,,,, itieasiired iri t.he cores. There is excellent agreement. For example, from 1,440-- 1,450 ft in zone E, both curves indicate 40 % water saturation. This is also the estimated S,, value tor the zone sirice little displacwrient of heavy oil by niud filtrate is expected.

LC'ith S,, eqiial to S,,,, \vater resistivity can be calculated from ,4rr:hie's equation. Rwrranging that eqiiation gives

R,, y= El, . (S,, . $/e)' (5.39)

Intl ieinter\d 1,340-1,450ft, It,- 250olirn-m,+ = 0.37,s- = 0.4, andc = 0.9 (for sands), which gives R, = 6.8 ohm-rn. A similar calculation for the lower part of zone D where R, is 10 gives R, = 0.9 ohm-rn. This shows the rapid variability of Rw. Many sands in this area are steani flooded, which contributes to \vi& variatiuri in R,.


sp 1 R S F L -100 0,3 1000

0 150 1 O 0 O 50 A C A L

- _ _ _ _ _

Fig. 5-56 EPT-derived porosities and water saturations in a shallow heavy- oil well [courtesy Schlurnberger and SPWLA)

189 POROSITY LOGS

In cases where wells are logged through intervals in which steam has broken through, the steam appears to the EPT tool as hydrocarbon gas

. However, steam

overlay. When suc Neutron-Density response and the S,,, value from the EPT log can be corrected according1y.j4

Quick-Look Hydrocarbon Indication

Fig. 5-57 shows schematically how the combination of Induction, Den- sity, Neutron, and EPT logs distinguish between fresh water, salt water, oil,

5-51. for example, shows t,) in shale to be 22 nsim. Correcting for attenuation (750 dbím) by Eq. 5.30 gives ti," = 18.6 nsím. Assuming t,, = 7.6 for shale (from Table 5-4, for a 50-50 matrix mixture of sand and dry colloid$) and tv, = 26.1, Eq. 5.32 gives for the calculated porosity of shale

(+EP)sh = (18.8 - 7.6)/(26.1 - '7.6) = 0.59

However, Density-Neutron indicates a shale porosity of about 0.25 (Fig. 5-54), so the EPT reading is abnormally high.

A simplistic explanation is to assume the bound water in shale has effectively a higher dielectric constant than free water. Eq. 5-28 can be solved for t,,f if (b is known. That is

t,f = [tp. - (1 - 4) t,,l/rp

Taking the value of 0.25 as the true shale porosity

(5.40)

(tpJSh = t18.6 - 0.75 x 7.6lf0.25 = 52 nslm


The effectiye dielectric constant of the clay bound water in the zone considered is then given by Eq. 5.31 as

cr = (tpi)2s,,/ll .l = 52'111.1 = 2-14

This is roughly three times the dielectric constant of flee water

POROSITIES FORMATION FLUID INDUCTION L O G FDC-CNL-EPT

30- PU -0

h rn í m'----) 5

GAS

OIL

FRESH WATER

SALT WATER

1

I

! ---TL

I I I I I I I I I

I I

I I I I I I I I

L-

Fig. 5-57 Quick-look hydrocarbon indicator (after Schlumberger)

POROSITY LOGS 191

E n : b e y d \33,-0 tl?C1-<s.j~uC'*C\13 (_12 \-L?5,5kvtdad23 Gtl

L h \ C 3 L & 5 \r,tc-'\L.;_\\UC_\Cr~%*, 3; $5 ',ti

1,ahoratory nieasiirenients have shown that the dielectric constant of water-saturated porous rock indeed increaes with clay content. The increase is astounding at frequencies les5 ihuri 1,000 Hz where values up to 10" ha\,? been measured on shaly cores, but it is less pronounced with iricrrasing frequency.,'" This hehavior has heen theoretically related to the concentration of thin, plate-like particles that signify clay content in the rock, íO,í7 The platelet effect is predicted to be small at the 1.1 gHz operating frt.qiieney of the EPT, but logs indicate i t is not negligible.

l 'he anonialoiisly high dielectric constant of shaly formations is clearly related to the suhstantial increase in signal attenuation that occurs simultaneously. If the two effects can be tied together in such a fashion that the spccific surface area of entrained clay can he derived it would materially aid iri the interpretation of shaly formations.

I c

SUMMARY

COMPENSATED DENSITY LOG Measures formation density b y sensing attenuation of gamma rays between source and pad-rnounted detector applied against borehole wall. Auxiliary aetector corrects for mud cake and rugosity. Density variation: 2.0-2.9 g/cc. Depth of penetration-4 in.; vertical bed resolution-3 ft; logging speed-1,800 ft/hr. Conversion of density to porosity is exact; requires knowledge of mairix and fluid densities, Accurate matrix densities must be input to obtain good porosity values at low porosity. When gas is present, correct fluid densities must be inpul for good porosities, particularly at high porosity.

1, best porosity device in shallow, uncompacted formations 2. effective porosity obtained, even with shale or clay

3, can be ruT! in empty hole

1 ~ rnust be run slowly 2. quality degraded iri very rough hole; short, sharp

Advantages

present

* Disadvantages

washouts are particularly bad


R,v is unknown but should be constant over the interval matrix density and/or velocity are not known at least a few water-bearing zones of different porosities in the

formations of interest are clean

Two types of crossplots are the FIingle plot and the Pickett plot. With the former, valuesof R,% and matrix density or velocity applicable to the interval can be derived, and \ alues of porosity and water saturation for individual levels can be read directly from the plot. With the latter, H,, can also be derived but not matrix density or velocity. Instead. the value of the cementation component, m , is obtained.

THE HiNGLE PLOT'.'

(6.2)

ming R, and c are ose slope depends o

of qb vs 1 1 6 yields water-bearing line (

established, other lines representing different values of S,, can readily be drawn on the chart and values of S,, for plotted points read b~ interpolation between the line?.

Fig. 6-1 i F an example. Porosity iiirreases linearl) on the horizontal axis but ierictix ity decreases in ver! non-linear fashion on the I ertical axis, it i s waled such that l iv R I is linear. I'orosity for the plot i\ beit ohtained from Neiitron-Densit> logs, arid R, is ohtained from deep Induction or Lateroloq readings corrected for iiivasion.

The procedure is to plot values of cp and R, for all lebels of intere5t in an interlial, as shown on Fig. 6-1. The interval qhoiild be limited to a few hundred feet in depth over which R,, should be essential11 constant. Those p i n t s falling at lowest resistivity for a given porosity represent thr 100 10 u ater-bearing lex els. An S,. = 1 line is established b y drawing a straight line through the pivot point (4 = O, H, = o") and the most northnesterly points on the plot. This line can also be laheled R,.

The eqiiation €or the S,\ = 1 line is. from Eq. 6.2

C L E A N F O R M A T I O N INTERPRETATION 201

R , is determined by substituting in this equation values of (b and R, for any point on the S, = 1 line. Taking (b = 0.10 and R, = 6.5 as illustrated and placing c = 1 .O, since the plot is for a carbonate section, we find R, = 0.065 ohm-m.

It is equivalent to averaging R, values calculated in all zones believed to be water bearing.

T hni on for es

íi' E E .c O

ü

/ x 3

4 E ' .-

6 :

> u) u)

.- u .-

:o i 20

50

1 O0

I O00

O 5 10 15 20 25 Porosity (O/.)-

Fig. 6- 1 Porositwesistivity crossplot (Hingle method)

202 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION C L E A N F O R M A T I O N INTERPRETATION 203

llavirig detrrriiined the R,, line, those linea representing value, of S, other than unity are established a5 follows. For a fixed porosity tlie Archie equation is

(6.4)

The S,& = 0.5 line, for example, is therefore rcpresentt:d by points wlitw R, = 4R,, for any given porosity. We pick a convenient porosity \vhere ii is easy to read R,,, in this case the point (4 = 17.5, Rc, = 2. l), and multiply R0 by 4 to obtain a point (4 = 17.5 , 13, = 8.4) for which S,, = 0.5. ‘The S, = 0.5 line is established by joining this point tu the pivot point (4 -= 0, R, = a) as illustrated. Other lines of constant S,, art: drawn similarly. li’or example, S, = 0.3requiresR, = 11.1 H,sothatthepoint (I$= 17.5, R, = 23) fixes this line.

With a grid of constant S,. lines s o establkhrd, the S, value corresponding to any plotted point can be in-iniediately estirriated. For example, point 7 has an S, of approximately 28 ?i .

This technique allows quick evahation of a large number of levels. Note that it is unnecessary to calculate H, to obtain S,, values.

The Sonic-Resistivity Crossplot

If a Neutron-Density combination is run, porosities can be derived without knowledge of lithology and plotted directly on the crossplot. I low- ever, if only a Sonic log is available for porosity (the Iiidi.i<.tiori-Soiiic conibination is still popular i n some areas) and thelithology is unknown, porosities cannot be initially determined. A different procedure is then followed.

The horizontal axis of the plot is scaled in psecift, increasing from lrft to riglit with the scale covering transit times, t, from those observed on t l i r log down to possible matrix values. Fig. 6-2 is an cxarnple, with a scale froni

Values oft and R, for levels of interest are then plotted and an S,, = 1 liiie is drawn tlirough tlie most northwesterly points. lcxtension of this line to H = determines the matrix travel time, in this case 55.5 psecíft, which i s also the zero porosity point. The porosity corresponding to the travel time value at the right hand edge of the plot can then be determined by tlie M’yllie relation, Eq. 5.16. In this case fort = 110, t,,, = 55.5, and ti = 189 (normal assumption for usual water), the corresponding porosity with no compaction correction is

50-110 psecíft.

d, . ( 1 10 - S5.5)/(189 - 55.5) = 41 Yo

The horizontal u i s is then scaled linearly in porosity between the 4 = 0 point arid the 4 = 41 % point If the travel tínie in adjacent shales, trh, is greater than 100 psecift, the porosity value so determined must be divided

0 ’ 6 o

I I I I I , t (p sec/ft)l -+, 1

50 55 60 65 70 75 80 85 90 95 100 105 O 10 2p 4) 30 40 -

0.6

0.7

0.8

0.9

1 .o

2

3 E

4 0 ri

E S

12 10 : I 15 20 30

60 1 O0 200 1 O00 32

FiQ. 6-2 Sonic-resistivity crossplot (Hingle method)


by the compaction correction factor, Bcp( = t,,,/100), before establishing the porosity scale.

From there the procedure is the same as described for determining R,

is case a ma indicatessandstone, soc = 0.9. Eq. 6.3 usingthepoint (# = 0.20, R, = 3.0) on the R, line, gives R, = 0.15 ohm-m.

The Density-Resistivity Crossplot

If only a Density log is available for porosity and matrix density is unknown, the procedure is analogous to that with Sonic logs. in this case the horizontal axis i s scaled in grams per cubic centimeter, increasing from right to left and covering observed values up to possible matrix densities. A typical

right edge to 2.85 g/cc at the left. Plotting the ing the S, = 1 line to

nt 6 = 0. Assumingp,, of the plot at p = 2.3

2.30)/(2.68 - 1.0) = 22.6%

The horizontal axis could then be scaled linearly between 6 = O and 22.6 % . Procedure from there on fo1lou.s that described for the Sonic caw.

?he Movable Oil Crossplot

In cases where a Microlaterolog or MicroSFL log has been run, normally in conjunction with a deep Latwolog rather than a deep Induction R. curve, the crossplot can be extended to determine movable oil. To do so the foregoing procedure is followed, plotting (+, R,) values, establishing the porosity scale (if necessary) and the S, = constant lines, and determining R,. It i s particularly advisabte to correct the LLll resistivity readings for invasion before plotting them as R, values since corrections can be as much as a factor of two. Fig. 6-3 is an example with plotted points shown as X.

Each level is then plotted again at its same porosity but at a resistivity value equal to RVLL (or RMSFL) for that level multiplied by the ratio R,,/R,,, whereR,, is the value on thelog heading corrected to the temperature at the zone of interest. This ratio normalizes the R,, values from the microresistiv- ity curve to what they would be if the water in the flushed zone had the same resistivity as that in the undisturbed region. On Fig. 6-3 the points so


o / / 110

o 2 4 6 8 10 12 I I

dJ P o i ___ - ,

Fig 6-3 Movable oil crossplot

obtained arc shown as circles. The S,,t.aliies read for these points are actually S,, values.

In water-bearing intervals the two points corresponding to a given l e \ d shoiild both fail close to the S,, = 1 line. Levels 13 and 14 are examples. In

206 ESSENJJALS OF MODERN OPEN-HOLE LOG INTERPRETATION

hk.drocarhon- bearing zones Lvhere the hydrocarbon is immobile, both points \tAI again fall close together but at an S,, valueless than unity. Points 9, 10, arid 12 are cases in point. Oil the other hand where movable hydrocarbon is present, the points of a given level will separate significantly. The difference in indicated S,, values represents the movable oil (Sxc,-S,,), as a fraction of the pore space. For example, point 8 indicates that movable oil constitutes approximately (75% -35%) or 40%) of tht. pore space at that level.

Additional Remarks Concerning the Hingle Plot

First a word of caution. Two types of crossplot blanks with different divisions of the resistivity scale are found in service company chart books. One is labeled F = lí+' or ni = 2. This can be used for either carbonate or sandstone formations, providing Eq. 6.3 is utilized for calculating R, with c = 1.0 for carbonates and 0.9 for sandstones. The resistivity scale can be divided or multiplied by 10 (or any factor) without changing the validity of the plot. The other plot is labeled F = 0.62/$'.'', which is valid only for sandstones. In this case the basic relation is

(6.5)

This equation, with S, = 1, must be used in determining I{,* rather than Eq. fj.3. Use of the In = 2 chart for all conditions is recommended; it is simpler.

The presence of gas can be detected (by crossover) and correct + arid S, values obtained with the Density-Neutron log combination. With Density or Sonic logs alone, however, gas is riot obvious, derived porosities \vil1 be too high, and S,< values wi l l be too low.

Similarly, in tlie case of carbonates, lithology can vary between lirne- stone and dolomite in the interval being analyzed if Density-Neutron logs are available but not if only Sonic or Density logs are run. In fact, even if the lithology varies between sand, limestone, and dolomite, the crossplot car1 be niade with D-N porosities and a value of 0.95 utilized for c in calculating R,. Derived water saturations will be reasonably correct. Crossplots made only with Density or Sonic logs will not be valid.

CLEAN FORMATION INTERPRETATION 207

M'lierr formations are shal!., the plot is iinreliablr cx rp t in limited areas such us tht. Gulf C«a;t wherc~ reasonable S>,, \.aliies ( h u t a l ~ ~ ~ o r ~ n a l l y high effecti\,e porosity valiies) may he obtained b y wing ikrisity-Neutron or Sonic porosity values. Use of Density porosities will give unciidy high \vater saturations.

THE PICKETT

2, is 'The gener;ilized form of the Archie equation. as indicated in Chapter

St,," = (a/r$"')(R,,/ll,) (6.6)

which of course rediices to Eq. 6.1 when the saturation exponent, 11, equals 2; tlic cenientation exponent, ni, equals 2; and tlie cementation constant, a , equals 1 .O for limestones and 0.81 for sandstones (eqiiivalerit to c = 1 or 0.9 since c = Y'.). Itearranging and taking logarithms

log HI= - 111 log d + log (aR,,) - n log S,, (6 .7 )

l'his equation shows that if a, R,v, 11, and S,, are constant, a plot of log R, vs log 4 yields a straight line whose slope is - ni. A plot of R, vs + on log-log paper is called a Pickett plot.

Fig. 6-4 is an example of ;i I'ickett plot for a limestone reef, with poirits plotted for three xvells penetrating the same interval. Guidelines for plotting are the same us for the Hingle plot except that porosity values must always be

, C L

Fig. 6-4 Porosity-resistivity crossplot (Pickett method)


plotted on the porosity scale, not bulk densities or travel times. After the points are plotted, the R, line (Su = 1) is fixed by drawing it through the lowest resistivity points corresponding to different porosities - in this case

Eq, 6.4 also shows that if the S, = 1 line is extrapolated to 4 = 1 (at which point both log 4 and log S, are zero). the R, value so intercepted is equal to aR,. In this case, aR,+ = 0.040. If R,, is unknown. a value for a must be assumed and R, derived. In this case. since formations are limestone, a can be taken ai 1 .O so that R, = O 040. However, if R, is accurately known from other sources such as measurements on produced water, its value can be inserted and a derived. For example, if samples showed R = 0.050, a would be 0.80.

for S , = 0.3, R, = 25R,, for S,- = 0.2, etc. With these lines wrater saturation can be quickly estimated for any plotted level.

Where only Density or Sonic logs are run, matris density or travel times must be assumed in order to arrive at porosities for crossplotting. i n this casp ,onid = 2.71 gicc is assumed. The matrix constants cannot readily Le determined from the plot. If improper values are assumed, porosities will be in error but water saturation values will be correct (providing the lithology i s constant). Remarks on the Híngle plot concerning gas and shaliness apply also to the Pickett plot.

The advantage of the Pickett plot over the Hingle plot is that it does not assume the standard values of u and m built into the Hingle plot. This is important in low-porosity formations where variations in m can alter S, values considerably. Consequently, the Pivkett plot i s preferable in hard rock areas. However in such cases Density-Neutron porositiesshould be used to avoid the uncertainties in matrix density or travel time encountered when only Density or Sonic logs are utilized.

CLEAN F O R M A T I O N INTERPRETATION 209

1 RANGE O F UNCERTAINTY IN CALCULATED WATER 1 SATURATIONS

in measurement of R,, R,, and 4. Many case studies in the literature have reported values of m and n for

specific formations. A considerable variation has been observed. Table 6-1 lists average values of m and n for various formations, as determined from core measurements in one laboratory.G For sandstone samples studied, m varies from 1.5 to 2.0 and n from 1.3 to 2.2. Thespread is significant though average values are not far from 2.0.

= 2.05 - #. However, the determining factor really appears to be clay content, at least for sands. On general grounds the saturation exponentn should be essentially equal to m. The constant c would remain at 1 .O, which it theoretically should be if the foregoing equations for m were adopted.

In the case of carbonates, the value of m is directly related to the fraction of pore space that is vugular - that consisting of separate vugs connected by intergranular porosity. Laboratory measurements have shown that m increases from 2.0 to 3.0 as the fraction of vugular porespace increases from O to 60% . I 2 The effect is important at low porosities (5-10 % ) where water saturations calculated using the usual m = 2 value can be much too low.

Uncertainties in the measurements of R,, R,, and # are in the range of f 5-10 % . When these are coupled with uncertainties of the same order of magnitude in m, n, and c , the net result is an avf;rap uncertpinty in calculated water saturation in the range rt 10-20% .I3 For example, S, = 0.40 means S, = 0.40 k 0.04 at best and perhaps kO.08 at worst.

Maximum precision in S, is needed for reserve calculations or for deci- sions on whether to complete wells when S, falls in the critical 45-65% range. Consequently, it is desirable to do everything possible to minimize uncertainty in S,, such as measuring R, on water samples, checking log

21 o ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION

pormities against core porosities, making borehole and bed--thickriess vor- rcvtions to the logs d i e n necessary, cross-checking logs on adjacent wells, and deriving ni and n valiies from I'ickett plots. E\,eri so, reducing the

TABLE 6-1 V A L U E S OF m AND n F O R V A R I O U S F O R M A T I O N S

Lithology Ave m Ave n ___ -___---

Wilcox, Gulf Coast Sparta, So La. (Opelousus) Cockfield. So. Louisiana Government Wells, So Texas Frio. Su Texas

Miocene, So. Texas

Travis Peak and Coiton Valley Rodecsa, East Texas Fdwards. So Texas Woodbine, East Texas Annona, No Louisiana Nacatoch, Arkansas Ellenburger, W Texas Ordovician Simpson. W. Texas and New

Pennsylvanicm. W. Texas Permian, W Texas Simpson. Kansas Penrisylvanian, Oklahoma Bortlesvilie. Kansas Mississippian. Illinois Mississippiari, illinois Penrisyivaniari, Illinois Mudison, No. Dakota Muddy, Nebraska Cretaceous, Saskatchewan, Canada Bradford. Pennsylvania Frio. Chocolate Bayou, Louisiana Frio, Agua D i k e , South Texas Frio. Edinburgh. Sourh Texas Frio, HGIIOW Tree, South Texas Jackson, Cole S d , South Texas Navarro, Olrnos, Delmonte. So. Texas Edwards time, Darst Creek Co. Viola. Bowie Field, No. Texas Lakota Sd Crook Co.. Wvomina

Mexico

ss ss ss ss sc

Cons SS Uncons SS

l iD CS LS LS ss

Chalk ss

LS and Dol ss

1s ss ss ss ss LS ss ss LS ss ss SC ss ss ss ss ss ss LS LS cs

1 9 1 9 1 8 1 7 1 8 1 95 1 6 1 8 2 0 2 0 2 0 2 0 1 9 2 0 1 6

1 9 1 8 1 7 5 1 8 2 0 1 9 1 8 1 8 I F 1 7 1 6 2 0

1 55-1 94 1 7 1 1 8 2

1 eo, 1 67 2 01 1 89

1 94,2 02 1 17 1 5 2

1 6 1 6 2 1 1 9 1 8 2 1 2 1 1 7 1 6

2 5 1 5 1 3 3 8 1 6

1 8 1 9 1 3 1 8 1 9 2 0 1 9 2 0 1 7 2 3 1 6 1 6

1 73-2 22 1 6 6

1 47 , l 52 1 6 4 , l 69

1 6 6 1 4 9

2 04,2 06 1 1 5 I 2 3

2 8

.. - Source, G R Coates and J L. Dumanoir, "A New A p p r o a c h to Log-Derived Permeability." SPVdiA logging S y m p o s i u m Tfiinsochons ( M u v 49731

CLEAN F O R M A T I O N INTERPRETATION 21 1

uncertaint) in S, b e i o ~ t 10 ?U of the calculated value is difficult, e5prciallv if shaliness is a factor.

I

I ____ __

M U L T I M I N E R A L I D E N T I F I C A T I O N

In this section we consider how the relative amounts of different minerals making up clean complex formations may be determined by combining Density, Neutron, Sonic, and Spectral Cnmma Ray data. The principal minerals of concern are sandstone, limestone, dolomite, and anhydrite. Perturbing components are shale and less-frequently encountered minerals such as gypsum, salt, polyhalite, and sulfur.

Identification of matrix makeup is particularly important in tight formations for several reasons. First, porosities may hover near cutoff values, about 570, so the most accurate values obtainable from logs are desired. Dolomite and shale, for example, cause similar separations between limestone-based Neutron and Density porosity curves, but effective porosity is computed differently in the two cases. With dolomite the effective porosity is close to the average of & and Qo; with shale the effective porosity is closer to &,. The GR curve may not distinguish between the two situations because dolomites frequently are radioactive.

Second, tight formations often require acidizing or acid fracturing to stimulate production. Optimization of this operation requires knowledge of the rock matrix.

The third reason is geologic in nature. The trend of matrix development across a field may indicate preferential directions for offset wells. For exaniple, dolornitization is often accompanied by increased porosity so that the direction of increasing dolomite content may be favorable for production.

There are three methods of combining information from the basic porosity t0dS to distinguish mineral composition. They are the M-N plot, the MI11 plot, and the Litho-Density-Neutron method. The first two of these utilize Density, Neutron, and Sonic data. The third uses Litho-Density and Neutron data; it can also make use of Spectral Gamma Ray information.

THE M-N PLOT

The hi-N plot is a means of combining data from the three porosity devices in such a manner that effects due to porosity variation are almost

i

21 2 ESSENTIALS OF M O D E R N OPEN-HOLE L O G lNTERPRETATlON

eliminated and those due to matrix changes are maxirni~ed.'~ The quantities M and N are defined by

M = 0.01 (tr - t)/(p, - pf) (6.8)

N =

where t, ph, and q5N are log values of Sonic travel time (psecift), bulk density (glcc), and Neutron porosity (limestone units, fractional); and t,, ,or. and& are the corresponding values for pore fluid. The latter are taken as 189, 1 .O, and 1 .O, respectively, for fresh mud and 185, 1.1, and 1 .O for salt mud.

Porosity variations simultaneously increase or decrease the numerators and denominators of M and N so the parameters are almost independent of porosity. On the other hand matrix variations cause M and N to change. The values for single-mineral formations are obtained by inserting appropriate

Two matrix triangles have been drawn for average fresh-mud condi- lions. They represent the common combinations of sandstone-calcite-dolomite and calcite-dolomite-anhydrite.

To utilize the chart, log values of t, p) , . and 4, are read at a level of interest, M and N values are computed. and the corresponding point i s plotted on the chart. If the formation consists of a binary mineral mixture, the point will fall an the line joining the two mineral point?. i f three minerals are present. it will fall within the triangle formed by the three mineral points. A point such as A could represent a mixture of either sandstone-calcite-dolomite, calcite-dolomite-anhydrite, or even a combination of all four minerals.

In principle, if the matrix is known to consist of only three components, the relative amounts of each can be determined by the position of the plotted point relative to the apex points of the appropriate triangle. However, the method is not accurate enough for this for several reasons:

0 the basic silica-calcite-dolomite points are not widely separated their positions are somewhat dependent on assumed matrix velocity

i

21 3 C L E A N F O R M A T I O N INTERPRETATION

1

the location of any log-derived point is subject to the statistical fluctuations in Density and Neutron logs I

P

;GYP SUM d

0.9

P5 f =o*< ,A $Y

4 4 SECONDARY POROSITY 5

1 1

0.6

\@ Schlumberger I 3.51

I

i __i

0.4 0.5 0.6 0.7 O. 8 O. 3

Fig. 6-5 M-N plot for mineral identification (CNL Neutron) (courtesy Schlumberger and SPWLAl

APPROX SHALE oFRESH MUD

REGION ( ~ t 1 O , 1, = 189)

.SALT MUD (Pi - 1 1. tc = 185)

N


indicated. Tlic plot is tlirrefore t i c s t used on a statisticd basis to indicate 1v1ici.c. I group o f poiiits from ii g i \ w log iri t tr id tend to cluster.

Fig. (;-G slio\vs (~s;iiiiplt~s Iroiii tlir Pcriiiiaii hasiri area of \!'est T t ~ a s \r,here loii -porosity. coiiiples-carI,oii;ite sequences are common. (The inatrix poiiits are those applic:at>ltx to the earlier SKI' Side\\.all Ntvtron tool. ) Plot (u ) indicatcs a stbctiori c011ip0~et1 primaril!, of t \vo birtar!. mixtures, doloniitti-calcite aiid doloiiiite-ariti!,c~ritr. v i th minor secoiidnry porosity. Plot ( b ) slio\vs calcite-doloinitc with no anliydritta along n.ith considerable secondary porosity that increases with dolornitc: content. Plot (c), from a shallow zone, clearl>, inclicates the presence of g!pwni arid ~oiiie shaliiiess in a basically calcite-doloiriite-aiiliy(~rite niistiire. Plot ( t l ) shows various amounts of shale i n a saiidstoiie-liniestone rnixtui-e.

Creation oi 11-N plots is a laborious process t)!. hand h i t is \vel1 adapted to log processing in a computer ccnter. In that e r i \ ironriirnt i t is of coniider- able assistance in zoning logs for aiitoiiintic interpretation. The sanie is true of MID and LDN plots.

THE MID PLOT

The Xi113 (niatris identification) plot is another rnetliot! of iitilizirig the sanie log data as for the 11-N plot. I-' It is based on determining the apparent niatrix density. ( Q , : , ~ ) ~ , aiitl tlie apparent rnatris tra\.el time, (tlt,',)A, for a level of intclret; plotting the \dues on the hlID plot of I;ig. 6-7: and observing \vlierc tlie point falls relative to the positioiis of the single-mineral points. The latter arc 1oc:ited at their ki ion . i i matrix densities arid t r a \ d times.

To deteriiiiiie t l ie ;rppart,nt iiiatris density at a gi\wi depth level, v2liic.s

of p i , and 4, are read from tlie appropriate Density-Neutron crossplot (Fig. (5-8). 0 1 1 this chart 1int.s of coristant (p,,,,), have been created b y interpolating bet\veen aiid extrapolating from the kno\rn matrix densities of sandstone, liiiiestone, and doloniitc. 'lhe ( p , l , ~ ~ ) ~ v a l ~ i e for an!' plotted point can be read from these lines. For vsariiple, pi, = 2.45 gicc arid = 0.20 qivcs

i3etrriiiination of apparent matrix velocit!. is similar. The travel time valiie, t , is read froin the Sonic log and inserted along with & i r i the appropriate Sonic-Neutron crossplot (Fig. 6-9). On this plot lines of constant (t,),,), have been created by interpolation arid extrapolation from the known clean matrix values. For t = 70 pcsecift and +x = 0.20, (t,,,), is 46 pseclft .

( p , , , , ) , = 2.76 gicc.

C L E A N F O R M A T I O N INTERPRETATION 21 5

o 90

0 8 0 : DOL

o 7 0 . . . . . . .

: M

0 6 0 . . . . . .

0 5 0 ; . . . .

040 (a)

GYP

. . . . . . . . . . . . . .

0 50 O 60

,

N 0 5 0 ; . . : I , . \ . , . . . . . : . . . .

0 4 0 ( C ) 0 % 0.60

090 : . .

O 8 ( i ; . '

0 7 0 : . .

: M

O60 : . .

O 5 0 , .

.o. . . . . . . . .

. . .

: N : . . . . . . . . . . . . . . .

O 60 0 4 0 (b) 050

100: . . . . . . . . . . . . . . . . . . . .

o y " . . . . . . . . . . . . . . . . . . . . . . . . . .

0 y0 . . . . . . . . . .

0 ?,l. . . . . . . . . .

: M

0 . . . . . . . . . .

N f 0 5@. . . . . . . . . . . . . . . . . . . . .

o 4 0 ( d ) o 5 0 o 00

Fig. 6-6 Examples cf M-N plots (courtesy Schlumberger and SPWLA)


Thecoordinate point [ (pn,Jd, (t,,,Jd] so obtained is plotted on Fig. 6-7 and the matrix components are interpreted as for the M-N plot. Point B is the example case which indicates a calcite-doiomite mixture.

The 9 u art a N plot. In particular, shale piis and them upward and to the right. No single point i s very significant by itself. Grouping of points indicates trends.

Salt ,, 2.0

2.1

MID* PLOT 2.2

2.3

( M A T R I X I D E N T i F l C A T l O N P L O T )

2.41

2.5 I E

5 2.6 \

O I E 2.7

2.8

Q v

2.91

I 3.0 \

f

3. i L

30 40 50 60 70 c i , , le , p s / f t

Fig 6-7 MID plot for mineral identification (CNL Neutron) (courtesy Schlurnberger and CPWLfi)

21 7 CLEAN F O R M A T I O N INTERPRETATION

The main advantages of the MID plot over the M-N plot are that no computations are required, the matrix points on the chart (Fig. 6-7) are not dependent on porosity or salinity, and the parameters plotted have a

als-salt, gypsum, and sulfur.

THE LITHO-DENSITY-NEUTRON METHOD

This i s a relatively new method of mineral identification.'' Its input is the bulk density, pt,, and the photoelectric absorption coefficient, P,, from the Litho-Density log, and porosity, &, from the Neutron log. The p, and Y , values alone can be used to derive lithology (and porosity) when only two minerals are present, as described in chapter 5. With the addition of Neu- tron porosity, three minerals can be distinguished.

The method is based on a plot of the apparen versus the apparent volumetric absorption index, shown in Fig. 6-10.

is derived from its p,, and P, values as The volumetric photoelectric absorption index, U, of a given formation

(6.10)

Thiscoefficient is additive for different components in a mixtureso that for a formation of porosity 4

(6 11) u = 4 * u, -t (1 - +,)u,,, where U, is the absorption index for pore fluid and U,, i s the absorption index for the matrix. Values of these parameters are listed in Table 5-2.

For a formation of unknown matrix, rearranging Eq. 6.11 gives the apparent matrix absorption index as

(Un,,), = (U - Uí # ) I ( ' - 4) (6.12)

The procedure for a level of interest i s to read the log values of &, pb, and P, and insert the first two in Fig. 6-8 (or equivalent) to find apparent matrix density, (p,,),, and apparent total porosity, 4+a. For example, 4, = 0.20 and Pb = 2.52 give (p,,),= 2.8 and #,, = 0.16.

are inserted in the nomogram of Fig. 6-11 to find This nomogram solles Eqs. 6.10 and 6.12. The example indicated with P, = 3.65 gives = 10.9.

Next, the values of Y,, pb, and


The \ d u e s of (p,,,), and (Uiiir).so obtained are inserted in the plot of Fig. 6-10. Approximate percentages of limestone, sandstone, and dolomite can be read from the grid lines. The example, point C indicates almost equal amounts of dolomite and calcite with poscibl>, a small amount of quartz.

Fig. 6-12 shows example plots from various zones in Lvells of mixed lithology and low porosity. The groupings of points clearly define different lithologies. Some intervals contain gas, as indicated by points shifted up\ \wd from their true locations. Gas causes (p,,,,), to decrease without significantly altering (UZ,,Ja.

~ C N L NEUTRON POROSITY INDEX. P u

Fig. 6-8 Chart for determining apparent matrix density (courtesy Cch lum berger)

C L E A N F O R M A T I O N INTERPRETATION 21 9

I O 0

90

80 * r

\ yi

i ul

'I t- 70 t- v) z [L t-

-

a

9 z 60 O v>

'L1I

50

40

/

O IO 20 30 40

4c.4~ NEUTRON POROSITY INDEX, p u

Fig. 6-9 Chart for determinirig apparent matrix travel time (courtesy Schlumberger)


2.70}

2.75

2.80 ~

2.85

2.90 I

fication (2)

,

9 10 11

Fig. 6-10 p-U plot for mineral identification (courtesy Schlumberger anu SPWLA)

_1

I 6 5 4 3 2 1 4 6 8 1 0 1 2 1 4

Fig. 6-1 1 Nomogram for determining apparent matrix photoelectric absorption coefficient (courtesy Schlumberger and SPWLA)

221 CLEAN F O R M A T I O N INTERPRETATION

As with the M-N and MID plots, the presence of clay or shale is a very perturbing factor. Fig. 6- 13 shows approximate locations of kaolinite, illite, and chlorite points as well as that of feldspar, which frequently

O

D

a a IO ir 14 boa

Fig. 6-12 Examples of p-U plots (courtesy Schlumberger and SPWiA)

p CHERTY LIMESTONE

D


acconipaIiir,s cla!,s. The locations of i inh~dritc~ and salt points are also sho\\~ri .

Cia‘. C J ~ shale iricltision shifts points ton.nrcl the l o \ r w right, caiisinh quartz forniatioiis to appear cxccssi\.cly dolomitic and calcite-dolomite inter\ als to appear antiydritic. 'The amhigiiit>, c w i lie resolved to scme extciit \i.iih Spectral Camiiia Ra!. data using Fig. (i 11. This is an empirical plot of percent potassium versus ppm thorium on n,hich the approximate locations of (Th, K) points for formatioris containiiig 100 ' ' i , lo\v-potassiiini clay sucli as kaolinite (Cl,), 100 % high-potassium clay such as illite (U), and 100 % feldspar (Fel) are indicated. Log-derived points are plotted on thischart? and the percentagesof the two different typesof clay and feldspar are estimated. Corrections can then be made to lithologies indicated on the

Fig. 6-15 is an example of a shaly carlionate zo~ie'. At first glance thepU plot indicates primarily a calcite-doloinite-ati1iydrite mixtiire. Iíowever, the tliorii.iin-potassiiIiri plot sho\vs the majority of levels are slialy and contain a Inistiire of iiigh-potassium clair and feldspar. \4'lien this is takm into

(P,,,,), - (U,,,,), Plot.

2 . 4

- 2 . 5

E

>. 2 . 6 t

D

9 .-

v) z 2 . 7

5 U

I I- z w 2 . 9 a- 4 n n .Z 3 . 0

5 2 . 8

K - F F L D S P A R

A N H Y D R I T E

0 KAOLINITE . ILLITE CHLORITE

3 . 1

A P P A R E N T MATRIX C R O S S SECTION (Umaa)

Fig. 6-13 p-U plot showing clay arid evaporite locations (courtesy Schlumberger and SPWIA)

CLEAN F O R M A T I O N INTERPRETATION 223

Constant percentage by weight lines - - - - - - Feldspar

C' 1

L c12

--- - p I' . 2 5

Fig. 6-14 Thorium-potassium plot for determining clay and feldspar content (courtesy Schlurnberger and SPWLA)

account, it indicates the matrix is composed of variable amounts of calcite, clay, and feldspar with little dolomite and no anhydrite.

TRENDS IN M U L T I M I N E R A L I D E N T I F I C A T I O N

[Ve can expect to see further refinement of the Litho-Density-Neutron method. The p-U plot is better than the 11-N or MID plots because the basic quartz-calcite-doloiriite-anhydrite points are more widely separated and there is no uncertainty in matrix travel time to contend with.

Even with the addition of Spectral CR information, however, there is not enough inclepcndent data to di:;tingiiish the various matrix components iirianibiguously. The induced Gamma Spectroscopy tool, * a new logging tool currently k i n g tested, will do much to close the gap. This tool gerierates bursts of neutrons that interact with the formation nuclei and cause them to emit gamma rays of characteristic energies. By detecting these gamniaays and cataloging their energies, the relative amounts of hydrogen, calcium, silicon, chlorine, sulfur, iron, carbon, arid oxygen can be determined arid,

' Terinwl CST (Canirria Spectroscopy 'r(J0i) hy Schlumberger and C/O by Dresser-Atlas.


2.6.

2 7 -

2 8

2.9

3.0

(Pm,),

from them, percentages of sand, limestone, clay, and other components can be

At present, existing tools in this nature are being directed toward locat- ing hydrocarbons behind casing, where the greatest need exists: but there is

Q

-

1

-

r . Salt

jn/ 2.5 A

3 1- o 5 10 15 20

Th

J

K ~~

Fig. 6-15 Mineral identification using both p-U and Th-K plots (courtesy Schlumberger, O SPE-AIME)

i i

i i

1

i I I t


no reason they cannot be used for mineral identification in open hole. Combination of the Litho-Density, Spectral Gamma Ray, and induced Gamma Spectroscopy logs will resolve most mineral components in thenear future.

REFERENCES ‘A.T. Ilingle. “The Use of Logs in Exploration Problems,” SEG paper, (Los

’W.H. Fertl, ‘“ingle Crossplot Speeds Long-Interval Evaluation,” OGJ

’R.H. Lindley, “Use of Differential Sonic-Resistivity Plots to Find Movable

“G.R. Pickett, “A Comparison of Current Techniqueí for Determination of

y Cross-Plotting,’’ SFWLA I,ogging S y n i -

r , “A New Approach to Log-Derived y r n p o s i z m Transactions (May 1973). , “The Influence of Particleshapeon the Forma-

tion Resistivity Factor of Sandstones and Shales,’’ SPE 1560-G (Denver: October

‘P.N. Sen, “The Dielectric and Coiiductivity Response of Sedimentary Rocks,” SPE 9379 (Dallas: September 1980).

“A.E. Biisrian, “.4 Generalized Archie Equation,” SPWLA Logging Sympo- siirtn Transactioti~ (July 1982).

‘”J. Clemenceau Raiga, “The Cementation Exponent in the Formation Factor Porosity Relation. The Effect of Permeabilitv,” S F \ V L A Logging Sytnporirirn Tranr- aciions (June 1957).

“D. K. Sethi. “Some ConFiderations Aboiit the Formation Resistivit) Factor - - Porosity Relations,.’ SPWLA Logging Symposium Transactions (June 1979).

“F. J. Lucia, “Petrophysical Parameters Estimated from Visual Descriptions of Carbonate Hocks: A Firid Classification of Carbonate Pore Space,” Jozir. Pet. Tcch. (March 1983), pp. 628-637.

‘’C. Khelil. “Analysis of Errors in Logging Parameters and Their Effects on Calculating \Vater Saturation.” SPWLA Logging Syrnposicm Trurisocfions (May

“J.A. Rurke, R.L. Campbell Jr.. and A.W. Schmidt, ‘“The Litho-Porosity Crossplot.” The Log Analyst (November-December 1969).

”C. Clavier and D . H Rust, “The MID Plot: A New Lithology Technique,” The Log Arialyrt (IVo.r,ember-Deceniber 1976).

“J.V. Crues Jr . , “Lithology Crossplots: Applications in an Evaporite Ba\in-the ?via\ erick Basin of Southwest Texas,” SPWLA Logging Sytnposiurn Trarisactions (June 1917).

Angeles: 1959).

(Janiiar! 1919). pp. 113-1 18.

Oil in Permian Formations,” J. Pet. Tech. Vol. 13, No. 8 (August 1961).

Water Saturation from Logs,” SPE paper (Denver: 1966)

1960).

1971).


1; J.S. Oardner arid J.L. Durn~iiioir, “Litho-Deriait), Log Irit(:rI>retatioii,” S P R ‘ I A Logging Syrtiposin?fl Trnrisucfioris (July 1980).

I5 1’. Westatvay, R . IIerzog, and H.E. Plasek, “The Garriiria Sprctrometrr Tool-Inelastic and Captiire Garririia Ray Spectroscopy for Reservoir Analysis,” S P E 946‘1 (Dalla\: Septenil~er 1!180).

1Y IZ’.A. Gilchrist J r . , J . A . Quirein, Y. L. B«iiteniy, and J .R. Tahanori, ”Appli- cation of Canirna Ray Spectroscopy to Formation E:valuation,” S1’\iJ/Lr\ Logging Syrnposiitrn i’rutisuctions (July 1982).

SHALY F O R M A T I O N INTERPRETATION l i e presence of slialc in resermir rock is an exireniely perturbing T factor in formation evaluiition. On tliex one hand it coriiplicatec the

I determination of hydrocarbons i n place: on the other hand it affects tlie ability of thr reservoir to produce those hydrocarbons. Most sands contain some shale or clay. The effect of this is to

reduce the effective poiosity, often signficantly lowrr tlie permeability, soinetirnes drastically alter the resistivity from that predicted t q r Archie’s equation *

Clay. which is a major component of shale, consists of extremely fine particles that have very liigh surface area and are therefore capable of binding a substantial fraction of pore water to their surfaces. This water contributes to the electrical conductivity of the sand but not its hydraulic conductivity. It cannot be displawd h y hydrocarbons and will not flolv. For this reason we define ejjwtiue poro,~itiy as the port: space occupied by only rionclay-bound fluid and total porosi fy as that occupied by both clay-bound arid noricla>r-bound fluid.

A shnly 1iydrocarl)on-bearirig formation can exhibit a resistivity little different from that of a nearby clean water sand or that of adjacent shales. This nieans shaly pa)’ sands can be difficult to find on resistivity logs and, e\wi if they can, application of the standard Archie equation can give water saturations that are too pessimistic.

A case in point is shown in Fig. 7-1 . I At first glance none of tlie sands \voiild be picked us productive on the Induction log. A close examination kind shaly sand anal’ , however, shows that the interval niarked PERF contains an estimated 14 7’0 iiydrocarboris by volunie. After perforation, the zone produced approximately 161,000 bL1 oil, 61,000 Mcf gas, and 114,000 bbl water in the first two years. This production could easily have been overlooked.

Too rniich shale in a reservoir rock will kill its production through excessive reduction in permeability. However, a modest amount of shale, if

227

1 j


disseminated in the pores, can be beneficial in trapping interstitial water and permitting commercial hydrocarbon production from zones of abnormally high water saturation. The foregoing example is a good illustration of this situation. If the perforated zone had been a clean, coarse sand of the

I N D U C T I O N - E L E C T R I C A L LOG

R E S I S T I V I T Y

O ohms 1(

Fig. 7- 1 Low-resistivity pay sand (courtesy Schlumberger)

229 SHALY F O R M A T I O N INTERPRETATION

water-oil ratio, though at greater rates. However, the fine clay particles in the pore space increased the irreducible water saturation and allowed production at the actual 0.7:l water-oil ratio.

7-2 function of the conductivity of the saturating water, C,.

If the sand is clean, the plot will be a straight line passing through the origin with slope 1íF as predicted by the formation factor relation, Eq. 2.1. * Writing it in terms of conductivity (the inverse of resistivity)

(7.1) C, = C,/F = 'p2 * C,

However, if some of the rock matrix is replaced by shale, maintaining the same effective porosity, the line will be displaced upward and the straight portion will interrupt the C , axis at some value, C,,,. This is the excess conductivity contributed by the shaliness. It follows that use of the Archie saturation equation

(7.2) s, = c m / 4

t CO

Fig. 7-2 Excess conductivity contributed by clay (courtesy Schlumberger) ~~

'This statement is not entirely true. Even clean sand grains have a small surface conductance thatgivesC,afinitevaluewhen C,= O. Thiseffectisimportantwhenloggingfreshwaterweils but is insignificant in normal hydrocarbon logging situations.


will gi\.e saturations that are too large because C, will be abriorriially high for a given effective porosity, 4. A modified water saturation ecluation that includes a shaliness term must bt! used.

THE NATURE O F SHALE

Shale is a mixture of clay minerals and silt laid down in a very low-eriergy environment, principally hy settlemelit from still water. Silt consists of fine particles, mostly silica, with sinall amounts of carbonates and other nonclay minerals. The solids of a t)rpical shale may consist of about 50 % clay, 25 % silica, 10 ‘70 feldspar, 10% carbonates, 3% ironoxide, 1% organic material, and 1% other material. The shale may also contain 2-40 ‘% water by volume. It is the clay component of the shale that affects logs in abnormal wa!x

Clay is comprised of crystalline clay minerals.’ These are hydrous aliinii- nuni silicates of the general formula S(Al,O,) . Y(Si0,) . Z(OH), which contain small amounts of other elements such as magnesium, potassium, iron, and titanium. Clay of detrital nature is a weathering product of preexisting rock, so its composition is quite variable, depending on the environment and conditions of trmperature, hiiniidity, and acidity under which it was formed.

Clay particles have a layered platelet structure. The crystalline platelets are very thin, 5-10 A, but may extend to about 10,000 A in length or width. They are stacked one above the other with spacings between them of 20-100 A. The clay particles are therefore eutreniely small about 2, in inaxirnum dimension.’ This is 10 to 100 tinies smaller than average sand grains. Therefore, there is ample space in sandstone pores for clay to exist, and indeed it is found there.

Clay minerals are classified into specific groiips according to their crystal-structure. Those of concern in sedimentar!, rocks are montrnoril- lonite (a form of smectite), illite, kaolinite, chlorite. and mixed-layer minerals. Table 7-1 lists properties of these clay groiip that arc’ important iii

formatiun evaluation. The first data column gives an important parameter, the cation

exchange capacity (CEC:). Note that niontriiorilloriite arid illite have much larger values than chl»rite and kaolinite.

The second colunin lists the porosity that the CNL Neutron log N W I I ~ read theoretically in a 100 O/o -dry clay formation because of hydrogen bound in the crystal l a t t i ~ e . ~ ‘l’his hydrogen does not contribute to conductivity.

’One angstrom u n i t (A) equais i ~ ~ c r n ; one micron ( p ) equals 10-‘cm. __I

-

SHALY F O R M A T I O N INTERPRETATION 231

TABLE 7 - 1 CLAY PROPERTIES OF CONCERN IN LOGGING

Spectral GR e3 h0<.1 Components i chcln 92

(’L“, h, íavi CEC p(av), Minor K, U , Th,

Clay Type meqig <bcNL gicc Constituents 96 ppm ppm

Montrnoril!onite 0.8--1 5 0.24 2.45 Co, Mg, Fe 0.16 2-5 14-24 Illite O 1-0.4 0.24 2.65 K, Mg, F e , T i 4 5 1.5 <2 Chlorite Kaolinite

__~-____________-_.__-_

Mg, Fe 0-0.1 0.51 2.8 - - - - 0.42 1.5-3 6-19 __ 003-0 06 0.36 2.65 ____--

_I__-

h.loritmorillonite and illite ha1.e smaller values than chlorite and kaolinite--the opposite of CEC.

The next coluriin lists the average dry cia!, density. It varies both with hydrogen concentration and with content of minor-constituent heavy minerals such as iron (succeeding column). There is a range on tlie order of +10% in both +c.N1. and p values since clays vary widely in detailed composition.

The final three colunins list average concentrations of naturally radioactive components in the clay.‘ Of interest is the high potassium concentration of illite and the high thorium content of montmorillonite.

Montmorillonite is somewha.t unique in that it swells in contact with water. \Vater intrudes between the platelets and forces them apart. The fresher the \vater, the greater the swelling. In fact, lattice spacing increases as (21 + 1 I / x ’ r ) A, where C is the water salinity in moiesiliter (1 nioleiliter = 60,000 ppin).’

Another feature of nioritmorillciiiite is that it undergoes diagenesis to illite at tlie higher subsurface teniperatures. This frees water and contributes to overpressuring of adjacent sands.

SHALE OR CLAY DISTRIBUTION IN SHALY SANDS

hlost logging tools average formation response over 2- to 4-ft vertical intervals. In these “iiriresolvable” interval5 shale or clay may be disposed in the sand in three ways or in combinations thereof: laminated, dispersed, alid structural (Fig. 7 3).‘-

Laminated

In this form, thin shale larriinations -- fractions of an inch to many inches in thickness - are interspersed with clean sand. The effective poros-


ity and the permeability of the shale are essentially zero so that the overall porosity and permeability (horizontal) of the averaged interval is reduced in proportion to the fractional volume of shale. For example, 40% shale will theoretically reduce effective porosity and permeability to 60 % of the clean

It would appear that laminated shale over 50% by volume could be tolerated in an otherwise porous sand. However, it is unlikely that thin shale laminations extend uniformly very far from the wellbore. The interlayered sand laminations may pinch out nearby. Consequently, 30-40 70 laminated shale is the maximum amount normally tolerable for production.

sa

CLEAN LAMINAR STRUCTURAL DISPERSED SAND SHALE SHALE SHALE

DISPERSED MATRIX

INTERCALATED LAMINA

DETRITAL ' MICA

RIP-UP CLASTS

INFtLTRATlON RESIDUES

Fig. 7-3 Forms of shale distribution in sediments (after Wilson, Schlurnberger)


The clay in Iarninated shale is of detrital origin. Since such clay is derived from diverse rock types and soils, it is generally mixtures of two or more clay minerals. Following deposition, the shale and sand laminations may tend to

shale laminations to have somewhat the same composition as nearby thick shale beds,

Dispersed

111 thi.; form clay, not shale, is disseminated in the pore space of the sand. It replaces pore fluid. This type of distribution is very damaging because a relatively small amount of clay can choke pores and reduce effective porosity and particularly permeabiiity to nonproducíble values. hlaximiirn

grows in place afte between the pore flu

result of chemical interaction ts of the sand such as feldspar.

o discrete-particle. pore- servations of scanning

electron microscope (SEhI) pictures (Fig. 7-4) .' The discrete-particle type consists mainly of kaolinite, 1% hich builds up as isolated booklets that lower porosity or permeability only a littie. The pore-lining type coatq the grains \kith Lvhiqkers, forming nricropore\ that trap a good deal oí' pore water and significantly lower pernic,abilit) . The pore-bridging type chokes the pore space with a mass o f tmdriis that significantly lowers Pffectivc port)sit!, and drastically rediice\ permcdiiity. Fig. 7-5 diows the effect on perineabilit!.. Pore-lining clay lowers permeabilit!. one order of magnitude: pore-bridging clay lowers it yet another order.

Because of their in situ origin, authigenic clays tend to be purer, more crystalline, and more likely to consist of a single mineral than detrital clays. Their composition may differ radically from that of detrital clay in nearby shale beds. In addition these clays are not subjected to overburden presiiire. so a ghen quantity of clay (particularly montmorillonite) may tra water than it would in a compacted shale. Thus, it is risky in log an assume that the clay in a slialy sand has the same characteristics as nearby shale, although we are forced to do so at present.


Structural In this form cla!~ grains, which may be aggregates of clay particles or

mudstone clasts. take the place of sand grains. Porosity and permeability of the sand is affected very little. Conseqiiently, this tyI)e of clay is least objectionable, hut it does not occur frequently.

_ _ _ _ - ~ --__- - FIGURE - O - "DISCRETE PARTICLE" KAOLINITE ---

FIGURE b . - "PORE-LINING" CHLORITE _.__

FIGURE C. - "PORE-BRIDGING" ILL ITE

Fig. 7-4 Forms of authigenic clay in sandstone pore space (after Neashan, O SPE-AIME)


1 000

500

1 O 0

50

n E i 10 J < C i w u

E a.

2

a

0 Q l o

O 5

O 1

o o5 n

6 10 14 18 22 26 30

POROSITY 'o

Fig 7-5 Effect of autbgeplc clay type on permeability (after Neachan, O SPE-AIME]

236 ESSENTIALS OF M O D E R N OPEN-HOLE LOG INTERPRETATION SHALY F O R M A T I O N INTERPRETATION 237

However, the particular disposition strongly affects the producibility of the formation. Dispersed clay is much worse than laminated shale. For example, a zone may calculate 15% effective porosity with 50% water saturation. Such a zone could be a 30% porosity sand interlayered 50-50

half-filled with authigenic clay. The former would produce at a much higher rate but probably with greater water cut than the Iatter.

tions. %y sand wr ace

SHALY SAND INTERPRETATION MODELS

Interpretation of shaly sands is still evolving. Over the years a large number of different models for calculating water saturation have been used. A comparative evaluation can be found e l~ewhere .~

Three methods once employed extensively are still used to some extent. They are oriented to logging tools available at the time and therefore may still be needed on older logs

1. The a - ~ p m a t i ~ c ~ - ~ ~ ~ ” n s . a t ~ o n _od (1950s) in which Sonic porosity and Induction resistivity are used directly in the Archie equation with

CATION EXCHANGE CAPACITY The most important property of clay in log evaluation is its cation

exchange rapacity. This i s the source of the excess conductivity depicted in Fig. 7-2.

tutions in the lattice and broken bonds at the edges. Charge-balancing cations, typically Na + , reside on the surface of dry clay. When the clay is in contact with a saline solution, these Na + cations are held in suspension close to the clay surface and, as a result, repel the C1- anions in the solution from the clay surface.

clay e n uit of

Thecurrent picture of the Ka+ ion and H,O molecule concentration near the clay surface is shown in Fig. 7-6.’’ Directly on the surface of the clay is a

milliequivalents per gram of dry le 7-1 shows that CEC is high for

mpensating effects. It is m- to high-pprosity san

t works best in_-Eg-

2. The dispersed model (1960s) using Sonic and Density porosities. The former reads essentially total porosity and the latter reads effective porosity in disperscd-ciay sands so that the difference is indicative of the degree of shaliness. The method is directed toward sands with authigenic clay biit has also given good results with laminated shale.

3 . The Simandous model (1970s) which uses porosity from Density- Neutron and shale fraction determined from GR, SY, or other shale indicators. This method has been the backbone of the service companies’ shaly sand interpretation programs for the last ten years. It is applicableio_disgersed ~r lamjnateci sha!_e,

At the mesent time a transformation to shaiv sand models based on

WATES

O U T E R H E L M H O L T Z

P i A N F

\4 S C H EM A T I C

WATER M O L E C U L E

L - .- cation exchange capacity, CEC, rather than on shale fraction, Vsh, is underway. TWO versions are in existence: the Waxman-Smits and the Dual- Water.

7-6 Water bound to a Clay surface (courtesy Schlumberger, @ SPE- AIME)

238 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION SHALY F O R M A T I O N INTERPRETATION 239

riiontmorillonite, intermediate for illite, and low for chlorite arid kaolinite. The high end of the range for each type probably applies to larriinated clay and the low end applies to dispersed clay hecaiise the former undergoes severe mechanical stressing during compaction that creates broken bonds.

Cation exchange capacity may also be espressed as milliequivalents per unit \.olunie of pore fluid, Q, given by

( 7 . 3 )

tvliere p is the density in gicc of the dry clay particles arid 4 is the porosity of the clay.

Relation of CEC to Surface Area

Q = CEC . p ( 1 --4) i+ iiie(1icc

Measurements have shown that cation exchange capacity is essentially a reflection of the specific surface area of a clay regardless of type.” This is depicted in Fig. 7-7, which shows that a single value of approxiniately 450 ni2 of surface area per riieq of cations applies to the clays of interest. (This implies that the negative charge density at the clay surface is the same for all clays.) Surface areas of clays. varying from 800 d i g for niontniorillonite to 20 m2/g for kaolinite, are hundreds of tinies greater than those of sands, which vary from 0.01 to 5 ni’lig.

Clay-Bound Water The model of Fig. 7-6 predicts a very importarit clay property, i.e.,

there is a layer of water next to the clay surface that is essentially immovable.

Fig. 7-7 Relation of area to CEL for API standard clays (after Potcheti and SPWLA)

The elwtrostatic hincliiig force is s o strong that the water cannot be squeezed out ~ v e i i by trernenclous overburden pressures. By the sanie token, oil niigrating into a shaly sand n,ill not replace this water.

The amount of claybound or anion-free water has been measured on siialy sand cores and the results are @\-en, in grains of water per meq of esctiangeaLle cations, as12

\ic’ = 0.22 + 0.084l\~c: (YA)

\vhere C is theconcentration of sodium chloride (molesiliter) in the water in eqiiilibriiim. Assuming the densit!, of the bound water is close to 1 . O , Eq. 7.4 also represents the volume of hourid water in ccimeq. Dividing W by the spcxcific surface ;irea rif clay, 450 x 10‘ cm‘iineq, gives the approximate thickness of the bound la;,w as 4.9 + 1.9 v‘¿

For a formation that has a cation exchange capacity of Q riieqicc of pore fluid, this l t d s to the fundamental relation that the fraction of pore water, S,, that is bound to the cia!. is

_-

S,, = \V . Q (7 .5 )

(2 values of produciblr: slial? sands range tip about 1.0 nieqicc. i f ’ is qq)roxiriiatel>~ 0.3 cc,’rineq, ivhicli 1neari.s that iip to about 30 t7c of the pore water can be bound in such sands. Shal~-santls with Q v:iliies higher than 1 .O are generally too tight to produce,

A shaly water-bearing sand therefore contains tivo types of water: bound water tied to the (slay and free water in the remaining pore space. ‘The latter is really not all free, since it includes irriidiicible water associated v i th the wnd grains, but i s eyiiivalent to the water i n a clean sand.

Shaly Sand Partitioning

Fig. 7-8 shows the partitioriirig of a hydrocart)ori-bearing slialy sand as erivisioned with this model. Tlie rock matrix is comprised of normal sand particles, silt particles, and clry clay particles. The fluid is composed of hound water, free water. and hydrocarbons.

Tlie total porosity (hound water + free water +hydrocarbons) is designated as qbtz and the free or effective porosit!’ is 4,. The latter is given by

4, = +,(l-S,) (7.6)

The volumetric fraction of hydrocarbons is

+t, = 4 4 - S%t) (7 .7 )


where S,&, is the fraction of total pore space containing water. This is a difficult quantity to determine in shaly sand interpretation. The charge- balancing Naf cations (also called counterions) that are associated with the clay give rise to electrical conductivity. This takes the form of cations

imposed. Externally it manifests itself as the excess conductivity illustrated in Fig. 7-2.

The manner in which the counterion conductivity is considered to act in the pore space affects the calcidation of the total sand conducti1,ity and the evaluation of water saturatioil. Two models are 113 current use: the Waxman-Sinits and the Dual Water.

nother when an

Sand

L 1

Solid

Fig. 7-8 Partitioning of a shaly sand in Dual-Water model

Waxman-Smits Model (W-S)

normal sodium chloride electrolyte

water conductivity at 100% ?at

In this conception the cation conduction and the conduction of the

where C,* is the conductivity of the free pore water in mhoím and B is the specific counterion conductivity in mhoim per meqicc.

LVhen hydrocarbons enter the pore space and displace free \vater, the counterions are more concentrated in the remaining water; the effective water conductivity becomes

(7.9) C,, = C," + BQIS,,

e con- g expression

G = is,, * #+)'(L.+ BQ/S,J* (7.10)

This equation can be solved for S &, and C , are obtainable from the and Q to be determined.

The value of R has been measured by applying Eq. Ti. io to shaly sand cores with known Q values saturated with water of different salinities. B is 3283 mhoim per rneqicc and at water salinities greater than

temperat& dependent, rising to 25-at_3iiOs0F. " Application of Eq. 7.10 then hinges,on the determination of Q. This

matter will be deferred until later. One major objection has been raised to the 1%'-S model. It predicts,

through Eq. 7.8, that water sands of constant C,, but increasing shaliness uilI have increasing effccti1.e water c«ndiictivities to the point that shales should appear to contain quite saline water. There is a good deal of evidence to the contrary. The Dual-Water model was devised to circimi\vent this constraint.

30,OQO ppm. A t lower salinities it increases somewhat. It is also extremely I

- I

'Stricti) cpraking. t h k equation \h«iild be w riitrn

C, = S,," . 6t"'(C, + nQ/s,,)/a

For siniplicit~ the ~atiirat~oti arid cementation exponents n and t n habe been taken a\ 2 and the cementatioti coiirtatit o 1

z

0

L

u

(3 O i

Ly

0 I 2 8 z y5 n

O E

8

5 2 ,


Since shales are laid down in waters with moderate salinities. the value of W is approximately 0.30, giving QSh = 3.3. Consequently, all well-compacted shales should have Q values of approximately 3.3, regardless of the particu-

Shale Porosity Consider a shale in which the fraction of solids that is dry clay is Y and the

cation exchange capacity of the clay is CEC,, meqig. hssuming that grain densities of dry clay and silt are the same, the CEC of the dry shale is

CEC,h = CEC,I + Y

Applying Eq. 7.3 with Q = 3.3 and p = 2 65, leads to a prediction of shale porosity as

(7.19)

" .- - - - - - _ _ I

Porosities calculated from Es. 7.20 are listed in Table 7-2 for the range

TABLE 7-2 CALCULATED SHALE POROSITiES

4 sh

Average CEC (meqlgm) Y = 0.35 Y = 0.55 Y = 0.75 _- _- Clay Type

¡hie O 25 o O7 o 10 O 13 Kaolinitekhlorite O 04 o o1 0015 O02

Montmorillonite 1 00 O 22 o 30 o 37

Table 7-2 shows that shale porosities can vary widely. The greater the fraction of clay and the greater its CEG. the higher the shale porosity. Allowing for the range of CEC values given in Table 7-1, montmorillonite shales will have porosities in the range of 20-50 % , illite shales will have porosities in the range of 5-20 % , and kaoliniteichlorite shales will have porosities in the range of 1-5%. This explains why shale porosities do not necessarily decrease uniformly with depth and why high shale porosities can be found at great depths.

The Density log will read shale porosities correctly, provided the grain and bound-water densities of the shale are the same as those for sands. normally 2.65 and 1.0 gicc. There is some evidence that shale values are


typically higher, perhaps 2.85 and 1.2 gicc, in which case the Density porosity will be a little too low. In any case it should be a clue to the type of clay in a shale.

TABLE 7-3 TYPICAL NEUTRON POROSITIES O F SHALES

tron porosities in shales are not clay distinctive. f 14-50 l o . In fact, Neutron tools using thermal

nt) may read even higher porosities if ermal neutron absorbers such as boron,

lithium, cadmium, gadolinium, and samarium in the shales. Although compacted shales can have high water contents, their effective

porosities are zero because all of the water is tightly bound to the clay surface. By the same tokcri permeabilities are virtually nil, in the range of 10 '-10 "d.

Shale Conductivity

Since compacted shale contains only bound water, its conductivity is, by Archie's principle

c,, = 4\h2 . Cb (7.21)

where & is given by Eq. 7.20 and C , is given by Eq. 7.14. The Dual-Water model therefore predicts that shale conductivities

should vary in accordance with the CEC capacity and content of clay, which determines &,, and with temperature, which largely determines Co. Confirmation of this prediction is ilIustrated in Fig. 7-9, which shows calculated versus measured shale conductivities for a wide range of wells. Published data listed log-derived Conductivities along with CEC,, and Y values from core analy~is.'~Porosities wcre calculated from the CEC,, and Y

246 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION SHALY F O R M A T I O N INTERPRET'ATION 247

1 J u e s iising E(l '7. 20. YiiLlished tcniperature5 perrnittrd cdlciilatiori of Ch using the tt.mlwrature dependeric) '"

Ct, = 6 8(1 + O 0545t - 1 127 10-'t2) i; 27) where t = ( "C - 25). The cerneritation exponent was allowed to \ ai > froni

Approximate porosity

1000 I --r--T----r I I 1 1 I I 1 I 1

1,000

' 500 v)

I

o - 3 u "a, 8 s 0 2 -

m o

n

O

1 O0 Corlductivity

Y

Overpressurec

/" "A

-1111111 1.000 3.c

Observed, inlihos

Two Louisiana wells A Two Wyoming wells 0 Eleven Oklahoma wells O One California well

Fig . 7-9 Calculated vs observed shale conductivities (D-W model)

1.8-2.2 ir1 accordance with 111 = 1.8 + 0.6 CEC,,,, since increased clay surfaw area implies increased tortuosity. Conductivities so calculated are compared to measured shale conductivities in Fig. 7-9.

The agreement between calculated arid measured values is remarkable, considering that the comparison covers 16 wells from Louisiana to Califor- nia with depths ranging from 2,000-15,000 ft, CEC,,, values ranging from 0.02-0.7, and temperatures ranging from 100-275°F. Some of the spread between nieasured and calculated values can certainly be attributed to logging tool averaging. CEC,, values for cores taken within 1-2 ft of each other varied by a factor of 1 . 4 , which corresponds to factor-of-two variations in calciilated conductivities.

The D-W niodel therefore explains why shale porosities can vary all the way from 2 % to 35 %G and resistivities can vary from 0 . 3 to 100 ohm-m. This being the case, it should work well for shaly sands.

APPLICATION O F THE DUAL-WATER METHOD T O SHALY SANDS

For practical application we shall use the following form of the D-iV saturation relation, derived fIorn Eq. 7.13 by replacing conductivities by resistivities (C = l i R ) and rearranging terms

S,~,tz - S,,, . S b ( l - R,,/H,J = R,/(R, * +r2) (7.23)

The second terni of this relation applies theshale correction. If it is omitted, the expression reverts to the familiar Archie relation. To apply the equation, tlie parameters SLi, $,, R,, and R, must be determined. First consider Sb, the bound water fraction in the shaly sand.

From Fkp. 7.5 arid 7.18, S, may be written as the following ratio (also called normalized Q")

To detrrrriirie S, accurately requires a direct measurement of Q; a Q-log is sorely needed. Unfortunately, no such log is currently available, although measurements on cores can be readily made in the laboratory or even at the wellsite.1Y,2" Consequently, we are forced into indirect methods using shale indicators.

In terms of Vrb, the volumetric fraction of shale (including its boiiiid water), the effectively porosity, $ e , can be written

4, = 4, - v,, dtd, (7.25)


where 4, is the total porosity of the shaly sand and 4t,h is the total porosity of the shale fraction in the sand, Equating this expression to that of Eq. 9.6 gives

7.26) ___ - - -

Determination of S, therefore reduces to obtaining Vsh from available shale indicators. This is a key result.

€valuation of V,,

No single logging measurement accurately measures V<,,. Consequently, V,, is usually estimated from several shale indicators and the lowest value is used."*'' The two best indicators are the Density-Neutron difference and

1. V, from the Density-Ne Because of the lattice-boun y, a gas-free shaly sand will

always read a higher Neutron porosity than Deniity porosity, as illustrated by the differences in Tables 7-2 and 7-3. The larger the fraction of shale, the greater the difference. The effect is linear, so the shale fraction is gibeii by

(V,h),rJ = (4% - Od)/(Gnsh - +d.ih) (7.27)

where the numerator represents the difference in Neutron and Density porosities in the shaly sand and the denominator represents the difference in nearby shale. The latter will typically be O. 15 to 0.30, depending on the amount and type of clay in the shale.

This method cannot be used when gas is present or suspected since gas distorts the 4,, and +a values.

2. V,, from The Gamma Ray Log Gamma Ray deflection increases with shale content of a formation.

Consequently, an index of the degree of shaliness of a sand is obtained by linearly interpolating between the clean sand level and the shale level

I,, = (GR - GR,,)/(GR,h - GR,,) (7.28)

1


where

GR

GR,, = average reading in nearby 100% shales, APIU

= reading in the sand of interest, APIU nu

I,,, will vary from zero in a clean sand to 1 .O in shale. Estimation of the clean sand and 100 70 shale levels, as illustrated in Fig.

7-10 is not always easy. There may be few clean sands and the shales ma!' vary considerably in activity. so a good deal of judgment is required. An occasional abnormally high shale reading should be ignored.

The fractional volume of shale, VSh, will be equal to the shale index, Ish, if the density of the formation does not vary with shale content. This is the situation when thin shale laminations are intermixed with clean sand layers

On the other hand when ing clay content is acc«mpanied by a as it is when authigenic clay grows in sitysands, then the curved line of Fig. Many cases will fall between the two

extremes. Consequently, a more generally applicable relation that might be applied when a Density log accompanies the Gamma Ray is

(V,d = I,, . (P /P \AJ (1.29) /-

where p i s the density of the formation of intereat and p,,, is the density of nearby shale. The exponent 3 is an educated guess: it has never been determined precisely.

U'here spectral Gamma Ray logs are run, improvement may be effected in certain arcas by eliminating the U component and determining (V&H using only the Th + K components. If feldspars or micaceous formationr are prominent, the potassium component should be eliminated or subdued.2' One field study reported excellent correlation between (Va,JGR and CEC values measured on cores and poor correlation between (V&,, and the same CEC values.25 This was attributed to the GR responding primarily to montmorillonate and illite, with high uranium arid potassium contents respectively, along with these clays having high CEC values. On the other hand the Neutron-Density separation gives greatest weight to kaolinite and chlorite, which have low CEC values. This is an argument in favor of the Gamma Ray as the better CEC indicator.

250 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INI'ERPRETATION

O 25 50 75 1 O0

I I I I I

Fig. 7-10 Determination of shale indication. Ish, from the GR curve

SHALY FORMATION INTERPRETATION 251

3. t',,, from tiir SI' Log Ir1 siniilar fashion, V,,, can be calculated from the SP log as

(V,,J,$> = (SP -- SPJ/(SI'\,, - SP,,) (7.30)

where ttie riumerator is the difference in rriillivolts hetween the SP level in the zone of interest arid the clean formation level and the denominator is ttie differenre betuveen the shale and clean levels (the SSP). This relation is \.alid only under certain conditions, as pointeci out in Chapter 3.

\2'ith sejveral V,,, \values so determined, standard procedure is to pick the lon.c>st value as the correct one, excluding the crossplot value when gas i s indicated. The reason is that most side effects cause calculated V,, values to

.2 .3 .4 .5 .6 .7 .0 .9 1 O o .1

Fig. 7-1 1 Conversion of GR shule indication, ish, to shale fraction. Vsh (courtesy Schlumberger, O SPE-AME)


be too high. Heavy minerals or neutron absorbers in the shaly sand will cause (Vsh)ND to be too large. Hole enlargements in the shales will cause (Vrh)CA to be too great; it is particularlv important to correct the GR readings

Consequently, the lowest value is picked, but even it is not likely to be very accurate.

Determination of Effective Porosity

The next step is to determine the effective porosity, de, of the shaly sand. The Density and Neutron porosities are first corrected for shale as follows

4 d c = 4 d - v+ ' 4 d s h (7.31)

4 n c = 4, - v,, * (7.32)

t, the corrected porosities should be close together. The art be taken as the average

4 e = (4dc + 6"$2 (7.33)

gas is present, it will show up as a crossover or enhanced c rrected porosities, 4,, being significantly less than &. This

reason for proceeding in this fashion. With gas the effective poro taken as

-_ 6 = G$"&? (7.34)

The effect of these calculations is illustrated in the crossplot of Fig. 7-12, which applies to sand or limestone provided the porosity values input correspond to the matrix chosen. On such a plot. clean formation points fall along the 45" line and shaly formation points fall to the right of the line. GaT- bearing formations will plot to the left if not too shaly. The shale point, S, may fail anywhere in the indicated shale zone, depending on the type and content of clay in the shale. Point P represents a gas-free shaly formation point. Correcting for shale translates this poiiit to P, (parallel to the line OS), and averaging porosities at P, gives the effective porosity, P2.

If the same sand contained gas, it would show up as some point such as P, on theplot. Corrertingforshalemovesthat point top,, accentuatingthegas effect. Correcting for gas via Eq. 7.34 is equivalent to translating point P, to the 45" line in a direction parallel to the gas correction line. This brings P, back essentially to Y,.


Determination of &,,, the total porosity of the shale, is required next. Unfortunately, there is no accurate method of measuring this quantity. Dry clay densities may vary all the way from 2.4-3.0 gícc, so the Density porosity t low equation is

+tsh = 4dsl i + (l - &) 4 n s h (7.35)

where-6-b-a constant between O L ~ a n ~ ~ l t O , depending on local experience. The total porosity, &, and the bound-water fraction, S,, for the shaly

sand are then

(7.36)

(7.37)

0.60

0.50

O 40

0 4

O 30

0.20

o 10

O

Direction of gas correction /

I

Shale zone

I I I I I

o 10 o 20 O 30 O 40 O 50 O hn

Fig. 7- 12 Correction for shale and gas effects

254 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION SHALY F O R M A T I O N INTERPRETATION 255

Evaluation of Water Resistivities I t rcrrinins to determine tlie free a d boiind-water resistivitirs i n order to

c a l d a t e \vater saturation. The free-\vat<:r resisti\,ity, R,, , is best obtairieil froni a riearby cleari-water sand. For siicli a sand, S,, = O and S,,, = I SCI that E(], 7.23 reduces to

(7.38)

where R,! arid are the observtid resistivity and porosity of the clean sand. An alternate method is to obtain R, froni the SI', a s described in Chapter 3.

Likewise, the boiincl-water resistivity, R,. is hest determined from a nearby shale. For shale, S, - i arid S,$[ = 1 s o that Eq. 7.23 gives

R,$ = H'I ' $ L , 2

Rl, = % . &r,li? (7.39)

where R,,, ar id cit,lb are the resistivity arid total porosit', of tlie shale.

Determination of Water Saturation

All of the factors required for water saturatiori deterinination throqgh Eq. 7.23 are thereby deterriiiiieci. 1'0 solve that eqiiation, it is useful first to calculate the apparent \vater resistivity of the shal'. sand, nhich is

R,,, = R, . (7.40)

Eq. 7.23 then can tie written

S,,~C - sirt . s,, (i - RJRJ -= r{,(/Ii ,,.,

c,vt = L + & 7 - ( 1 1 , \ '€Iaa)

(7.41)

frorn wliiclh the total water saturation is .

(7.42)

where

b = s, (1 - R*/€$,)/2 ( 7.43)

It is a re1atiw:ly simple matter to obtain S,,, valiies, provided an R,,~, log is recorded during logging \vith a Density-Neutron porosity input to the R,$, computer equivalent to Eq. 7.35. In this case R,, , R,,, and Rxtl for the zone of interest can be read directly from that log. Once S, is deterniiried, its value and those of NJR, and R,/R,,, can be entered into the charts of Fig. 7-13 and Su,t can be read as indicated by the dashed lines.

Total water saturations so calculated will be higher than effective water saturatioris that have been historically used. If desired, the latter can also be computed as

(7.44) s,,, = (Swt - Sb)/(1 - S,)

I s, = o

2 0

FIJ4,

1 .o

1 .o

0.5

O

O 1

.~ -- Fig. 7-4 3 Graphical determinat ion of total wate r saturation


Fig. 7-14 provides a graphical solution to this equation. Finally, hydrocarbon content as a fraction of total volume is

(7.45)

11 of the equations, shalysand interpretation is as much of an art as a science. The most critical parameter, Vch, is difficult to pick in many cases. The relation for total shale porosity, &,,, is by no means firmly established. There may be a lack of clean sands to establish R,. The applicable cementation and saturation exponents m and n may differ from 2.0, although the value of theeeinentation constant a i s immaterial when R,, R,, and R,, are determined as indicated. In short, there is no substitute for experience in the region of interest.

1.0 r I

O 0.2 0.4 S h

Fig. 7- 14 Graphical determination of effective water saturation

SHALV F O R M A T I O N INTERPRETATION 257

One final point: Values of dd and 4,, indicated on the log must be corrected to the appropriate matrix before calculations are performed. That is, if the logs have been recorded on limestone matrix but the shah forma- -

environmental corrections should be made before computation. Those that might be required are invasion corrections to Rt values, borehole size corrections to CR readings, and temperatureípressure adjustments to CNL values.

SUMMARY OF DUAL WATER INTERPRETATION

1 Read the resistivities, porosities, GR, and CP values in the sand of interest, in a nearby shale, and in a nearby clean sand. Correct porosity values to the appropriate matrix if necessary.

Convert Is% to V,, using Fig 7- 1 1

(vsh)SP = (sp - spcl)/(sp$h - spcl)

Choose the minimum value Omit (\/sh)Nii if gas is indicated.

3 Correct the porosdies for rhaliness

+)dC = 40 - v,, . %,P 4)"- = d, - v s , . +oct.

Look for gas indication (&(. < ddc)

4 Calculate the effective porosity of the shaly sand

No gas. 9, = (4oc 4- O n J 2

5 Determine the total porosity of the nearby shale

#Ish = f i +es, + (1 - fi)d>,,h

258 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION i 259 SHALY F O R M A T I O N INTERPRETATION

where

6 = 0.5 to 1.0

h. Determine the total porosity and the bound-water fraction of the sand

4, = q e + vsh ' d l s h

s a = v,,, . & h / 4

Rw = Re, . rbci2

7. Determine the free-water resistivityfrom a nearby clean sand

8. Determine the bound-water resistivity from a nearby shale

Rb = R s h ' $ish2

9 . Determine the apparent water resistivity in the shaly sand

Rw, = R, . 4t2 10. Determine the total water saturation corrected for shale

-- Sw+ = b -t a ( r ? ~ ? , ~ ]

where

b = Sb (1 - Rw/Rb)/2

11. Determine the effective water saturation of the shaly sand

s,, = (S, - SJ(1 - S,)

A, = 9+(1 - SW+l

12. Determine the volumetric fraction of the hydrocarbon

Example

Fig. 7--15 is a composite log from an offshore Louisiana well through a

and below this interval (not shown) are a number of clean gas sands with pronounced Neutron-Density crossovers (as much as 30 porosity units) and Induction resistivities as high as 30 ohm-m. What does the shaiy sand of interest contain?

massivesand-shaleseries. Thesand of interest is from 8,5054,545 f t . Above 1 1

Fig. 7-15

w o

~- Sha ly sand example-offshore Louisiana


Applying the above step-by-step procedure, gives the following results.

R, = 3 In the sandfrom 8,510-8,540ft

SP = -95

R S h = 1.2 4dSh = 0.20 4,,sh = 0.50 GñSh = 87

In nearby shale 8,470-8,500ft

SPSh = -75 i n nearby clean sands 8,178-8,193 f t and 9.180-9.193 f t (averages)

use gas is suspected.

(V,,),, = (-95 +122)/(-7.5 + 122) = 0.57 Choose VSh = 0.34

Calculation of ejfective porosity d>dr

+,, = 0.26 - 0.34 X 0.20 = 0.19 = 0.33 - 0.34 X0.50 = 0.16

Gas is indicated since qinc c: @bdr

4% = J<oiy2-+Tiqz = 0.17 Calculation of water saturation

&, = 0.7 x 0.20 + 0.3 x 0.50 0.29 (6 = 0.7) 4, = 0.17 + 0.34 X 0.29 = 0.27 S, = 0.34 X 0.2910.27 = 0.36 R,, = 0.65 x 0.34' = 0.075 R, = 1.2 X 0.2g2 = 0.10 R,", = 3 x 0.272 = 0.22 b = 0.36(1 - 0.075/0.10)/2 = 0.045 Su, = 0.045 + &49 + O.WS/O 22 = 0.63

SHALY F O R M A T I O N INTERPRETATION

S,, = (0.63 - 0.36)/(1 - 0.36) = 0.42 +h = 0.27(1 - 0.63) 0.10

261

effective porosity of 17 % and a water saturation in the effective pore space of 42 gó . Since that water is not tied to entrained clay but is associated only with the clean sand fraction, it is probable that the zone would produce considerable water with any gas. Consequently, this interval was not perforated. The clean gas zones below the interval (not shown) calculated approximated 10% water saturation and produced dry gas on production.

S U M M A R Y O F EARLIER SHALY S A N D INTERPRETATION M E T H O D S

Automatic Compensatio (1950s)

ethod, with only Resistivity and Sonic logs

Shale causes R, to read too low and 4% to read too high, compensating each other in the water saturation equation. However, observed porosity must be corrected for shale to obtain effective porosity. Relations are

\vh er e

o, = porosity from Sonic v ithout shale correction Ii. = resistivity from deep Induction

= lowest of CR and SP indicators $,5h = porosity from Sonic in adjacent shale

This method i s still used in the Gulf Coast to obtain quick answers.

by uncorrected D-N porosity given by Where Density-Neutron is run in place oi Sonic, 4, in Eq. (a) is replaced

Od" = + tPn2)12 (4 where

#<+ = porosity from Density without shale correction b,, = porosity from Neutron without shale correction

c)

9

(Y

h

.I

v

2 a

f I Q II 1

a

h

v

o

c= c

C

o

3

T3 94

La o

.I

u

E --.. .A

5, E '5 2 4 o

k 3

u-,

- O

3 u

vi c

c- 5' 3

v

%

2 3

- A

... .U

=I ai z -2 6 Ld A

c: 4-8

4

.I

i.

u

jr

.- 10 .... *, .I

.I

u a

c

+. - .I

i ,

w

.I

x

4-

VI

.* c O a


1 A

Fig. 7-17 Water saturation, Simandoux method, chart 2 (courtesy Dresser)


Dresser uses the S,, relation as indicated. Schlumberger uses the same expression with R, replaced by R,. (1 - V,,).

REFERENCES

‘M.P. Tixier, R . L . Morris, and J.G. Connell, “Log Evaluation of Low Resis- tivity Pay Sands in the Gulf Coast,” SPWLA Logging Sy~zposium Transactions (1968).

*R.E. Grim, Applied Clay hfirieralogy (New York: McGraw-Hili, 1962). ’H. Ednirindson and L.L. Haymer, “Radioactive Logging Parameters for

Common Minerals,” SPW‘LA Logging Syniposiuln Transaction9 (Juile 1979). ‘N. Huhovets and W.H. Fertl. “Digital Shaly Sand Analysir Rased on Wax-

m a n - h i t s Model and L.og-Derived Clay Typing,” SPWLA/SAID Logging Szlrtipo- siunz Transactions (France: 1981).

“W.H. Fertl and C.W. Hammack, “A Comparati te Look a t iYater Satiira- tion Computations in Shal> Pay Sands,” SI’WLrl Loggit~g Symposirrnz Trarttnciioiis (May 1971).

‘“C Clavier, C . Coates, and J Ihmarioir , “‘I he Theoretical anti Eicperirneli- tal Bases for the ’Ilia1 jt’ater’ hlodel for the Interpictation of Shal\ Sands“ ( D e ~ i er. October 9-12, 1977).

“J.G. Patchett. “An Investigation of Shale C<indticti\ it?,” SI%i’I,A Logging Styizpoiíwn Trarisacfiotis (Jtine 1973).

‘*Hill, Shirlej , and Klein, Ibid. J’M.H. Waxman and L . J . M . Sinits, “Elcctrical Conductivities in Oil-Bearing

Shaly Sands,” Sor. Pot. Eng. J . (June 1988). pp. 107-122. 1 4 M . H . Waxinan and E.C. Thomas. “Electrical Condiictivities in Oil-Rearing

Shaly Sands-I. The Relation Between Hydrocarbon Saturation and Resistivity Index; 11. The Teniperature Coefficient of Electrical Conductivity,” SPE jotrrnal, No. 14 (February 1974), pp. 213-225.

15 Clavier, Coates, and Dumanoir, Ibid. “Patchett, Ibid. ”Waxman and Thomas, Ibid.

266 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION Chapter 8

1. J iitiasz, ”Norrnalizrd Q, ---?‘tir, Key to Shal!. Satid Evaluation IJsirig The \,Vaxnian-Srriits E(1uatiori in the Abseriw of Core Data,” SI’WLA Loggirig Synipo- sium Trnrisurtiori~ í June 1981).

A.E. if’orthington, “An Automated hlethod for the Measiiremriit of Cation Excliarigr Capacity of Rocks,” Gcophy , No. 38 (Febriiary 1073), pp. 1.10- 153.

1.I.C. Thomas, “The rleterniinatiori of Q, from klembrarie Potential I l i a . ~urerrieiits o n Shal!. Sands,” 1. Pet. T w l i . (September Ir)7ti), pp. 1087-1U96.

“ A . I’oii~iori and R . Gayniard, “The Evaluation of Clay Content froni L . o p , ” SPii‘L4 L o g g i n g Syrriposizrrn 7’nirisactiotis (Xlay 1970).

--R.C. Ransoin, “Methods Based on Densit) Und Neutron \Vel1 L,i>gging Responses to Distinguish Characteristics of Shaly Sandstone Reservoir Hocks.” I ’h

Ir(

1’1

.![I

‘0.7

L o g Ariullysi, Vol 18, No. 3 (1977), p. 4 7 . 2.1c,c, _. a\ier et al.. .l. Pat. Tech. (June 1971).

G . híarett, P. Chevalier, P. Souiisite, and J . Siiau, “Shaly Sarid Evaliiatioii Usiiig Gamma Ray Spectrometry Apl)licd to the N o d i %a Jurassic,” Si’fb’fA Lug-

“\V.L. Johnsori and \\!.A. I,iiike, “Sonie Practicd Applications to Iriiprove Formation Evaluation of Saridstonvs ii i thr. klackenzie Delta,” CM’LS Logging Syniposíiim 7ransactions (Jurie 1976).

LbR.P. Alger. l J . L . Rayiiicr, Li’.R. Iioyle, and h1.P. Tixier, “Formation Den- sit) L.ug Applications in 1.iqiiid-Filled Holes,” J. Pet. 7‘edi. (March 1963).

P. Simandoiis. “l>it4ectric 3le;isiiremmts i i i Porous Media arid Application to Shaly Formations,” Revue de I’liistitiit Francais d u PetrCJk. Suppleineritary Issiie. i963, 1111. 193--215; Eiiglish translation in SPLVLA Reprint Voluiiie Shaly Sond (Jiily 1O82).

2d

g¿t?g St/?tL~líJY¡U?)l i ‘ r < i r i M < ’ f i í ~ ? I S (JLiILe 1976).

:i:

I

PREDICTION O F PRODUCIBILITY

S tandard rehisti\ it!, and porosity logs pro$& good aIis\\ws for the quantity o f oil or gas in situ but nut on the proc1ucit)ility of those hydrocarhis . Prediction of the latter requires kno\vledge of reser-

\vir pressure, iorniatiori pernivdbility, and irreducible \vater saturation. N o n e of these parameters is precisely deierniined t)> ndard logs, although under certain circumstances approxiiriiate values can he deduced.

The \vire-line tool that can be of considerable assistance in predicting prodiicti\4t!, is the Multiple Formation Tester. I t Inensures formation pressures, ailo\vs calculation of pernieability, arid retrieves a sample of reservoir fluid for analysis. \f’ith these paranieters a reasonable, though not irifalli- blc, estimate of production rate arid type can be niade.

Before discussirig the Formation Tester, we need to set forth basic f h v relations, re\riew some pertinent aspects of irreducible water saturation anti perrnwbility, and consider the limitations in estimation of permeability from lugs.

FLOW RELATIONS

The rate of production of oil, q,,, in stock tank barrels per day (stbld) frorn a homogenous formation i i i i d c ~ radial flow is iven to a close approxi- niation hy’ &&,\as’> k:b\ 5c*,otd+

ffC>p3 a i” FbWi Fy i ccr,-\* q1 = 1 o . 10-’k, . h (P, - P,)/(p, . 13) =I @YD (8 I )

$ 4 where md i L \Lac

r5L k, = h = height of tlie producirig interval, ft P, = shut in reqer\oir pressure, psia I.{ = pressure in tlie wellt~ore at the producing level, p i a p , , = oil viscosity at reservoir temperature, cp B = formation volume factor, res hblistb

For average oil (SO’API, COR = 500) the quantity pLB ir: approximately

rffectix e l)eriiieabilitl oí tlie formation to oil, rnd

unity so that

<lo = 1.0 . 10- k” . 11 (Pr - Pi) (8.2)

267

268 ESSENTIALS OF M O D E R N OPEN-HOLE LOG INTERPRETATION PREDICTION OF PRODUClBiLITY 269

Correspondingly, the rate of production of gas, qg(uM), under similar conditions is closely

qg = 0.10 k, * h (P: - P:)/(kL, * Z * T) (8.3)

k, = effective permeability of the formation to gas, md Z = gas deviation factor at reservoir temperature and pressure T = formation temperature, O R

For averagegas conditions the product p,ZT is approximately 5 so that

qg 0.020 k, . h (P,’ - P,?) (8 * 4)

the hydrostatic pressure of salt water at the depth of interest, that is, 0.46 psiift x vertical. depth in ft. However, this can be a substantial overestimate in old fields where zones may be depleted or a considerable underestimate where producing intervals are overpressured. Consequently, a direct measurement is desirable.

The bottom-hole flowing pressure, Pf, cannot be determined by measurement at the time of logging. It is a function of both reservoir parameters (pressure, water-oil ratio, gas-oil ratio, bubble point pressure) and well parameters (depth, perforation efficiency, tubing size, choke size and place- ment, flow-line size and length, separator pressure). In effect, when the well is opened the reservoir will increase its flow rate until the reservoir pressure, which is the driving force, is balanced by the back pressures consisting of the hydrostatic head of the fluid in the tubing, separator pressure, and friction pressure drops in perforations, tubing, chokes, and flow line. A systems analysis encompassing all of these parameters is necessary to optimize flow rate and to determine corresponding bottom-hole pressure, P,. Wellsite programs to perform such an analysis are becoming available from logging and testing service companies.

Barring a systems analysis, the reservoir potential may be characterized simply by its specific productivity index, SPI, defined as its production rate per foot of producing interval (h = 1) and per psi of pressure drawdown (P, -Pf = 1). For oil, Eq. 8.2 gives

(8 .5) SPI = 1 x lo-’ k, b/d per psi/ft

That is, the SPI of an average oil reservoir is its permeability in darcies, which is an easy relation to remember.

For gas it is more common to characterize the reservoir by its absolute open-flow potential (AOFP), obtained by placing P, = 0, in which case the specific AOFP per foot of producing interval is

(8.6) SAOFP = 0.020 k, * P: scfdíft

It i s clear from these relations that effective permeabilitv is the kev

one phase hinders the flow of the other phases. The reduction factor is termed the relative permeability, k,, which has values between one and zero. In other words, the effective permeability to a given phase is the product of the absolute permeability of the rock times the relative permeability of the phase.

Fig. 8-1 shows the relative permeability of water and oil in a typical rock as a function of water saturation of the rock.’ The left side of this plot represents the situation existing in the undisturbed zone of an oil-bearing reservoir well above the water table. Water saturation is at its irreducible value. S,,>. No water will flow, so the relative permeability to water, k,, is zero. Oil will flow virtually unhindered because the water exists only on the grain surfaces, at grain contacts, and in very fine pores, leaving all major passageways open for oil flow. Thus, the relative permeability of oil, k,*, is close to unity.

At the other extreme, the right-hand side of the plot applies to the invaded portion of an oil-bearing zone where residual oil occupies 10-40 % of the pore space and water occupies the remainder. The residual oil is immobile so that k,, is zero. However, water will not flow unhindered

270 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION PREDICTION O F PRODUClBlLlTY 27 1

because the resicliial oil is left as isolated globules occupying it number of tlie mediiini-to-large pore spaces. 'I'liese crihstantially reduce the number of branching passageways availatde to \vater flow arid thereby reduce k,, froin unity to a valtic in the range of 0.3-0.6.

Between the t\vo extrenies i s the situation thnt exists when oil and nrattxr flow simiiltarieously. Both k,,, and k,, are siibstaritially less than one, and iri

fact their s u n ia also significantly les5 than unity. In order to predict tlie rates at which water and oil \vil1 be produced from

a reservoir rock, we need to knou, the relative permeabilities at the existing water saturation as well as the absolute permeabilit!.. The effective permeabilities, k,, ar id k,, required for Eys. 8. 1 -8.ü are

obsolute

<' ,'+,hOt.;L

/-

c w , W A T E R SATURATION, F R A C T I O N S X O range f)jy, a l )~ \

F ig 8-1 Relative permeability to o11 and water in ari o11 bearing formation I (courtesy Schlumberger, O SPE-AIME)

$&- j;' ,<- ut it d42' \& klk 3,\b\ i z h'v) 1 ' --h: K,(,t f h r d

Empirical rclatioiis exist that allo\v estimation of relative permeabilities if both actual and irreducible water saturations are known. One of the simpler \,crsions for oil and \vater is"

k,,, = [@.9 ---S*)/(0.9 --S.,,J]?

k,,, = [(S, -S,J / ( l -S,,)I3

(8 .{I)

(8.10)

With the foregoing equations, the water cut---that fraction of liquid production that is water-can be derived. From Eq. 8.1, the water-oil ratio is derived as

Ii'ater cut is

(8 12)

Fig 8-2 shows w a t u cuts 90 calculated for light, medium, and heav) oils.5 ii'hen S,, = S,,,, no u ater flo\vs and IVC = O . Ti17hen S, > S,,, wine mater will be produced. The hea\.i<Ir the oil, the mort' reliictant i t i y to move and the greater the watpr cu t . The prohleni in appljzing these charts is that the valiie of S,,, is difficult to dpterniiric acciiratel! .

ivc = WOR/(WOR -t i)

IRREDUCIBLE WATER S A T U R A T I O N

'The \vater saturation profile in a thick, extremely lioniogeneous, \vater- wet, hydrocarbon-be:rrint! reservoir that has stabilized over geologic time \vil1 be as illustrated in Fig. 8-3a.fi Consider first curve U . Proceeding upward froni the free water level. water saturation remains close to unity for a few feet, drops rapidly over the next 50 ft in the so-called transition z r m , and finally 1eví.l~ off at a more-or-less constant value considered the irreducible saturation.

Two opposing forces are at work. Interfacial tension, Lvfiicfi may be thought of as an elastic skin between water and hydrocarbon, tends to hold water at grain contacts and in sniall pores. Its force is essentially proportional to l / r , where r is the radius of cunature of the water surface, closely related to grain and pore size. The smaller the grains and pores, the more tightly the water is held. Opposing that force is gravity tending to pull the heavier water below the lighter hydrocarbon. This force i s proportional to the height above the water table and the density difference between water and hydrocarbons.

-

Fig 8-2 Calculated water cut as a function of S, and Swi (courtesy SPWLA, reprinted in CchlumDerger)

At any given level, water remains in only those interstices small enough that capillary pressure created by interfacial tension balances the gravitational force. The equilibrium equation is'

1

40t ; = 0.6

Lagoonal SS Fine grained Shaly laminated

K = 140md a = 24%

CEC 2 3

- 400 - a;

8 - 300 '

P

d m .-. - O .o m c O m

C

.- I

200 5 I- al

\ lil ; Lagoonal carbonate

B nr.i..,- -A. -I- - \

B

O '*M: ~ Water saturation. oerrent nnw cnare

7 - -- . r- - -Y---

Fig 8-3 (a) Water saturation vs capillary pressure for several types of formations (courtesy Core Lab, 6 SPE-AIME)

r = smallest surface radius associatcd with \vater held at grain contacts or in pores, p

h = elevation above the free water table, f t p,, = water density, gicc

pi, = hydrocarbon density, glcc

where (8.13) The left-hand side is the capillary prtxssure and the right side is the gravita-

tional force, both in psi.

274 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION PREDICTION O F PRODUClBlLlTY 275

est pores. Thus, the effective water41 contact is slightly above the free water level. Above that point oil displaces water first in the large pores, then in the medium pores, arid finally i n small pores as elevation increases. The transition zone will be sharp if all pores are about the same size but more gradual if there i s a wide distribution of pore sizes (Fig. 8-3b).

At an elevation of about 10 ft, water has been displaced from all pores greater than about 10p in size. From that point upward, water saturation continues to decrease very slowly until only water electrostatically bound to the grains remains. This water is the true irreducible water saturation, S,¡. Its exact value is difficult to pinpoint because it is approached so gradually. However, the S , value 200 ft above the water table is generally close to SWi.

Beach SS Medium grained

80 A K = 2400 o F a 60-

o a, a

t;i 40

lu a w

..- C

(I> .- - - 3

E, u

20 -

C L_ I l l 1 I I I I I O

1 O0 10 1 0

Pore throat radius. microns

Fig. 8-3 (b) Pore entrance distributions for the same formations (courtesy Core Lab, O CPE-AIMEI

The saturation profile of curve B is typical of an average shaly sand with irreducible water saturation of about 30 B . Curve A illustrates the one extreme of an rilniost clean coarse-grained sand. l h e transition zone is much sharprr, and the irrediicible water saturation is aboiit 8 % . At the other extreme, curve C represents a sandy carbonate containing a large percentage' of clay. The transition is extremely gradual, and the irreducible water saturation is o \w 90 % . The correlation between clay content (indicated CXC \,dues) and irreducible saturation is evident.

Wien a reservoir is produced from some level abo\Je the w.ater-oi1 contact, the oil production is accompanied by pressure drawdown in that phase, which essentially adds to the gravitational pressure difftirence trying to remove uater from grain contacts. This plus friction effects between moving oil and stationary water will cause loosely held water to break away and flow ~vitli the oil. Water production \vil1 therefore be large if completion is in the transition zone but small if conipletion is in the irreducible

It is apparent that the relative amounts of hydrocarbons and water produced will he related to the difference between actual and irreducible saturation, S, -- S,,,. Stated another way, effective permeability is also a function of that difference. The criipirical equations 8.9 and 8.10 portrav

zo11<3.

.

The pertinent question i s how to dctermine S,,frorii logs. In the ideai45c;vv,

c' CV/{ case of a long (200-ft), homogeneous, hydrocarbon-bearing zone, there is no problem because the actual \\ ater saturation, 5,b, at the top of the sand i s close to the irreducible value, and this value of S,, can be applied through out the zone Such formations, however, are rare. More often the formation 15 short, 15 not homogeneous, arid docs not h a w a readiljr identifiable \vater table.

It is clear that S,,, cannot be les5 that S,,. the fraction of pore water bound by d a ) , as discused in chaptcr 7.

c,,, 2 s,, (8.13)

This nieans from Eq. 7.26

s,, 2 v,, . 4tdJ4, (8.15) I

1 or

s,, . 6 2 v,, 4th (8.16)

bníib i n c i t x a-, (05 j L ~ ~ 5 ------+ y o' c , / o t c 4 S t 3 I L ' ' 3 4% . I

276 ESSENTIALS OF M O D E R N O P E N H O L E LOG INTERPRETATION

where

6, = total porosity of the formation in question V,, = volumetric fraction of shale

The product Vsh&, can range from zero in perfectly clean formations to about O. 12 in very shaly formations.

To the above value of S,, should be added the irreducible water associated with clean matrix grains. Sand grains have a surface charge similar to that of clays, sufficient to attract electrostatically and immobilize about t u o molecular layers of water. The volumetric fraction of pore water so bound, S,@, will be dependent on the surface area of the sand grains. However, calculations indicate that the maximum value, even where a large portion of

explained ín chapter 7 for estimating Sb. Basically these rest on determining Vlh from shale indicators, on the assumption that clays in sands are of the same type as those in adjacent shales where the indicators are calibrated. One case study showed good correlation in individual formations between \',,from GR and SPlogsandS,,valuesmeasuredoncores, but thecorrelations differed between formations, implying the foregoing assumption is invalid .' This is to be expected with authigenic clays. Once again the need for a Q log from which Sb can be obtained directly is evidenced.

To predict relative permeability and water cut with any confidence, an accurate value of (S," - S,J is required. This is difficult to obtain from logs, given the imprecision in determining Vsh. In particular if V,, is overestimated, s,,,, will be overestimated, S,will be underestimated, and the difference will be magnified accordingly.

Empirical Method

A method of circumventing this difficulty to some extent is as follows. It has long been observed in the field that the product S,<@ tends to he constant in the irreducible zone of a reservoir. This is predicted by Eq. 8.16 if the

PREDICTION OF PRODUCIBILITY 277

shaliness of the interval is essentially constant. Consequently, if the product of actual water saturation, S,, and total porosity, Qt, is calculated for a

Fig. 8-4a is a plot of S,, against Q for a Wyoniingsand, and Fig. 8-41] is a similar plot for an East Texas carbonate.q In both cases most points conform

2.

2c

+ lis

'O1

I /

1 m d c = 0 .078

I i I 1 O0 O 20 40 s w 60 8 0

Fig. 8-4 (a) S, vs 9 for a Wyoming sand (courtesy Schlumberger and SPWLA)


to a h>~pert)olic line iiidicating S,,$ = constant. Those points arc corisitlercd to be in the irreducible water zones. The exception is the open circles on the cartmnate plot, Lvhich correspond to the bottom 10 ft of that fortiiatiori. These show higher S,S values and are interpreted to be i n the transition zone. Note that the constant is 0.078 for the sand and 0.008 for the cartmi- ate, indicating the fornier is much inore shaly than the latter.

If the Camina Ra). log or other shale indicators show that V,,, is not constant over the interval of interest, then the quantity (S , . +J/V,,, should be tested for constancy to determine if in fact S, represents S,,,. Howevrr,

O

Fig. 8-4 (b) S, vs @ for a n East Texas carbonate (courtesy Schlumberger and SPWLA)

PREDICT ION O F PRODUClBlLlTY 279

tliis introduces the iincertaint!, of V,,, deteriiiination arid rtduces the relia- hilit!. of the method.

In short, irreducible water saturation is a very important pararneter in productiori estimation, hut it is an eliisive quantity that is difficult to obtain from logs. ?'he one log that comes close tu defining S,,, is the Nucltaar Maqetisni (NML,) log. In sandstone it nieasures the aniount of fluid (of normal viscosity) free to move in the pore space. Therefore

where 4 = total porosity measured by Neutron-Density +,4ii = porosity indicated l>y KML

I n carbonates, however, the N X I L nieasures total porosity. The NMI, tool is not discussed i n this book because it is a very specialized clevice not yet coni In only r i i~ i .

E S T I M A T I O N O F PERMEABILITY F R O M LOGS

For intergranular-type formdtions, perrrieabilitv increases rapid! with porosity. However, permeability is also quite dependent on the surface area presented by the grains. A clayey, fine-grained formation wilt have a lower permeahility than a clean, coarse-grained formation of the same porosity. Flow paths dre finer and more tortuous iii the former case.

Empirical Relations

Irreducible \vatu saturation reflects surface area. Therefore, several empirical correlations have come into use relating absolute permeability to porosity and irreducihle \vater saturation for intergranular formations. A well-documented correlation is that of Timur, who madrcareful laboratory measurements on 155 sandstone cores from the Gulf Coast, Colorado, and California."' His correlations between porosity, irreducible water saturation, and permeability are shown in Fig, 8-5. With this data he derived the following relation

where k is in millidarcies, and + and S,, are fractional. This equation


0 PERMEABILITY IN M i L L i D 4 R C I E S

F i g 8-5 Permeability vs porosity and irreducible water saturation for various sands (after Timur, courtesy SPWLA)

I PREDICTION OF PRODUCiBlLlTY 281

predicted the measured core permeabilities to within a factor of two (standard deviation), which is quite good for permeability estimation. It should be noted that the porosities measured were total porosities, including bound water, if any. Also, irreducible water saturations were measured at

eliminate all but clay-bound water. An alternate cmpirical expression by Tixier i s often used"

(8.19) k = (250 4'/SW,)'

The two relations are plotted in Fig. 8-6a and 8-6b. For the same values of C$ and S,,, they give about the same answers except at extremes of permeability. A more elaborate relation gives similar answers. *'

80

70

60

50

40

30

20

I I I I I i I 5 10 15 20 25 30 35 40

O O

Porosity, I$(%)

Fig. 8-6 (a) Permeability from the Timur relation (courtesy Schlumberger)


60

- s? 2 50 v

c O .- c

L =I 40 u: 3 3 30

0

L 2 0 -

...

... m a,

o D

-

-

10

PREDICTION O F PRODUClBl l lTY 283

-

-

-

-

-

ln principal, these equations shoiild applj. equally well to oil or gas flow. Experience has shown, Iiowever, that k values obtained from the equatiorii; must be reduced by a factor of 3 to 10 when the expected production is gas. The reason for this is not clear, although partial explanations have been advanced.” One important factor rnay be that gas flow near the borehole is likely to be turbulent, whereas the equation applies to laminar flow. In any case the recomniencled expression for absolute permeability for gas flow is:

(8.20)

where p, is the gas density i n gícc and the bracketed term is obtainable frorn Fig. 8-ea.

Combining Eqs. 8.5, 7, 9, and 18, the specific productivity index (SI’I) for an average oil reservoir is approximately

k = p, (93 q’? ‘/S,,J2

k, Permeability (rnd) for oil ’O I

I l l I I I I I &I - 0 5 10 15 20 25 30 35 40

Porosity, + (%)

Fig. 8-6 (b) Permeability from t h e Tixier relation (courtesy Schlumberyer)

This is r in dppropriate relation if S,, and if S,,, are known

Using S, for S,,

Iri the more usual case where S,., cannot be reliably determined, the recommended approach is to tise the actual S, in Eq. 8.18 and csonsider the pcmneability so obtained as the effective rather than the absolute \ d u e . This leads to the simple expression

There is jristification for this procedure. At one extreme, when S,. = S,,,, Eqs. 8.11 and 8.22 are identical. At the otlier extreme, when S,, > S,,, they can givesimilai-values, dependirigon theniimbers. For example, if S , = 0.6 and S,,, = 0.3, answers are identical. This simply means that permeability calculated with C, instead of S,, can be a good approximation to effective permeability.

A similar expression for thespecific absolute open-flow potential of gas is

SAOFP = 170 p g . ($2 ’ is,)? . P,* scfdift (8.233)

T(J summarize, specific flonr potentials for intergranular íorniations, typically sandstones, can be calculated from log data to within a factor of about 3 if Sw, is reasonably well known. lVith the usual uncertainty in SWi, however, accuracy is probably no bettc,r than a factor of about 5 . Arid of course there is no wa)’ to take into account skin damage, if it exists.

For nonintergranular rock the above approach to productivity estinia- tion is not applicable. Formations containing solution channels, vugs, or fractures may have quite high permeability even though overall porosity is low, as indicated in Fig. 8-7. At the other extreme, formations with isolated pores such as chalks niay have high porosity hut quite low permeability. Both conditions are more prevalent in carbonates than in sandstones. For these situations one must resort to movable oil calculation or formation testing or localized log-core correlations. l 4


THE MULTIPLE FORMATION TESTER

The Multiple Formation Tester is a tool that can be run in open hole, positioned at a desired level, actuated to measure reservoir pressure and flow capability, and, if desired, further activated to take a 1-12-gal sample of reservoir fluid. *

*Designated Repeat Formation Tester (HFT) by Schiiimherger, Formation Multi-Tester (FMT) by Dresser Atlas, Multiset Tester by Welex, and Selective Formation Tester (SFT) by Gearhart.

285 PREDICTION OF PRODUCJBILITY

tlG?&mZ ?ceScJCefr - anydw-

Kekrcwer S o m p k s - 2 Give &mabi\i)ies- i '

Any number of pressure measurements can be made at various levels. lw6

Operation The Formation Tester has a hydraulically activated pad that goes

downhole in the retracted position. The tool is positioned at the desired level using a simultaneously run GR or SP log, and the pad is extended. This causes a small circular packer to seal against the mud cake on one side of the hole, as shown in Fig. 8-8. A metal snorkel tube about 0.5 in. in diameter,

FILTER PROBE

FILTER FILTER PROBE PISTON -

PROBE OPEN AND SAMPLING

PRESSURE GAUGE

CHAMBER b 1 íSLOVr RATE) EQUALIZING

CHAMBER # 2 (FAST RATE)

CH A %I B E A S

VALVE

(TO MUD COLUMN)

SEAL VALVE

(TO UPPER SAMPLE

SAMPLE CHAMBER1

SEAL VALVE

I T 0 LOWER

CHAMBER1

Fig. 8-8 Flow path in the Repeat Formation Tester (courtesy Schiumberger, O SPE-AIMEI


positioned in the center of the packer, is then forced through the niud cake into the formation. Simultaneously, a piston inside the tube retracts, uncovering filter holes and aliasing reservoir fluid to flow into the tool,

Reservoir fluid successive117 enters two small pretest chambers, each tia\4iig 10 cc of capacity (in ttie RFT). Flow is first into the upper chamber at a rate o f approximately 0.7 cclsec, determined b y the speed at which a piston i n that chamber is hydraiilically retracted. When that chamber is full, the lower one fills. Its piston is retracted t\vice as fast so the inflow rate is approximately 1.5 cclsec. Typically, both chambers fill in 15-30 sec, depending on the particular tool.

Pressure Recording

During tool setting and filling, prcssure at the probe entrance is continii- oiisly nionitored and recorded at the surface by means of the indicated pressure gauge. A typical pressure recording is shown in Fig. # - $ I . Time is proceeding from top to bottom at 6 sec per chart division. Pressure is recorded on an analog scale on the left-hand side for qualitative reading and in a digital format on the right-harid side for precise reading. The digital recording has a thousands, hundreds, tens, and units track. In each track the recording can be on only one of 10 discrete positions, either on one oí' ttie vertical chart divisions or halfway between. At level A, for example, pressure would be read as 4,351 psi. Thus, pressure can be read to 1 psi, which is the resolution of the normal pressure gauge. More recent recordings have pressure values to the nearest psi printed directljr in the depth column every few seconds.

Beginning at the top of Fig. 8-9, the pressure gauge reads hydrostatic, pressure of the mud (point A). 4t point R the pad sets against the rnud cake. Flow into the first test chamber starts at C and ends at E. Flowing pressure is read at D as 1,850 psi. Thesecond test chamber begins to fill at point E and i s full at point C . Flowing pressure is read at F as approximately 50 psi. At point G, the test chambers and flow lines in the tool are ful l and pressure builds back up to reservoir pressure. It is read at point H as 3,848 psi. This pressure is normally several hundred psi less than the hydrostatic pressure of the niud.

Pad Sealing

The pretest phase indicates quickly whether a good seal to the formation has been effected, whether the probe is tending to plug, or whether the

PREDICTION Ui- PKUUUC;IIILII Y L O I

Digiial pressure recording psi

looo Test N O __-

I 0-

1 O000

___.I o - - 4 1 O0 o -4

J

Fig. 8-9 Typical RFT pressure recording during pretest (courtesy Schlumberger, O SPE-AIME)

formation is completely tight. Figs. 8-10, 8-11 and 8-12 are examples of these problems. Poor seal is characterized by quick return to rnud hydrostatic pressure. Plugging appears as erratic flowing pressure. A tight formation is characterized during the first pretest by the flowing pressure dropping essentially to zero arid the chamber not filling in the usual 12-18

--

_I_---__-_

~i

ii

ii

i

Il

l

I

I ri-I

t 1 I t

290 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION PREDICTION OF PRODUClBlLlTY 291

gal sample of formation fluid withdrawn if desired. LVheri the sample chani1,er is filled or sampling is to be terminated, the seal valve is closed and the tool is unseated in readiness for repositioning or retrieving. Alternately, tlie other seal valve can be opened and the second chamber filled \villi reservoir fluid from the same l e \ d . The second sarnple should contain a smaller fraction of mud filtrate than the first, \vhicli should allow more accurate prediction of water-oil and gasoil ratios expected on production.

PERMEABILITY F R O M PRETEST DRAWDOWN I'ermeatiility may be calculated from the flow rates and pressure

drawdowns that occur while filling the pretest chambers. The applicable equation for drawdown pernieability, in niillidarcies, is

k = 5,050 qp1Ap (8.24)

where

5,650 = constant that takes into account tlie entrance diameter of the probe and the ciuasispherical flow of fluid into it. Thisconstant is for the HFT; an earlier constant was 3,300. For the FMT it is 2,760.

c1 = average flow rate during chamber filling, ccisec. Chamber \ d - Lime is knotvn and filling time is taken from the pressure recording

p = viscosity of tluid at reservoir conditions, cp. Intaken fluid is considered to be rniid filtrate, so p is normally taken as 0.5 cp

Ap = pressure drawdown, the difference between reservoir pressure (not rnud pressure) and flowing pressure measured near the end of tlie filling period, psi

This calculation must be made separately for each pretest chamber. For the recording of Fig. 8-9

Pretest chamber f l : q = 10110 = 0.625 ccisec A p = p = 0 . 5 ~ 1 )

3,848 - 1,850 = 1,098 psi

hence: k = 0.88 md

Ap = Pretest chamber #2: q = 101s = 1.25 cclsec

3,848 - 50 = 3,798 psi tierice: k = 0.93 md

Average drawdown permeability is therefore 0.9 md.

í>ra\\.do\vn prrnieahility repr rswts the effective pern~eahility to water i n t l i e in\xded zone. This muy represent only 30--509 of the atisolute iicmneahility xvlieri residual oil or gas is prewrit in that zone, as illustrated i n Fig. Further, the measurement has i w y shaiio\v depth of investigation -- o1i1y an inch or t\vo - since practically a l l of the prcAssure drawdo\vn occurs very close to the probe entrance. The ca1c:ulated value can therefore tx too 1o~v us a result of formation damageclose to the borehole, such as clay swelling. Sometimes formation cleanup during chaniber filling can actually b e obser\.etl, as illustrated in Fig. 8--13.

Because of these two factors, drawdowm permeability is often considered a lower liniit on absolute formation permeability. On the other hand, in consolidated formations the forced insertion of the probe can cause microcracks that stimulate rather than reduce permeability. This can lend to abnormally high permeability values.

Fig. 8-13 is a plot of Eq. 8-24 for various pretest flow rates. This can be used to make a quick estimate of permeability as soon a s the pretcst drawdo\i~n pressiire is seen. The range of permeabilities measurable is from approxiniately 0. 1-200 md. The upper limit is determined t q r the need to have at least 10 psi drawdown to obtain a reasonably accurate reading with a pressure recording of 1 psi resolution. h'ewersystems with resolution ofO. 1 psi increase the range to 2,000 nid. The lower liinit is achieved by waiting íor about 3 min for the pretest chambers to fill if flowing pressure drops to esseritiall). zero. In this case both chambers fill at the same forriiation- determined rate, calculated as total pretest chamber volume divided by total filling time, and only one pernieability calculation is made.

--

-- - - Drawdown decreases as formation cleans up -

- - - - - -

10.7 rnd I I 1 I I

I- Fig. 8-13 Example of formation cleaniip during pretest (courtesy Schlurnberger)

292 ESSENTIALS OF M O D E R N OPEN-HOLE LOG INTERPRETATION I PREDICTION OF PRODUClBlLlTY 293

1 O000

Ap. psi

Fig. 8-24 Permeability from RFT pretest drawdown

PERMEABILITY FROM PRESSURE BUILDUP

Following cessation of flow into the test chambers, the pore space in the \vicinity of the probe from which reservoir fluid has been withdrawn is repressurized by fluid flow from farther back in the reservoir. During this time, pressure at the probe builds back up to reservoir pressure. as illustrated by the G-H portion of Fig. 8-9. The greater the permeability, the faster the buildup. 17.1'~'g

Permeability can be determined from the rate of buildup. It is more difficult to calculate than drawdown permeability, but the value derived is not influenced by formation alteration near the borehole since the depth of investigation i s a few feet rather than a few inches.

Importance of Extended Measurement

After the second pretest chamber fills, following point G of Fig. 10-9, a small amount of flow continues into the RFT tool as the fluid in the flow

effect is called afterflow. The duration of afterflow, T,, depends very strongly on the compressibility of the fluid in the tool. It is given approximately, in seconds, by

T, = 10'Clk (8.25)

1 etet ber es un Pr* is

where

C = compressibility of the fluid, psi k = formation permeability, md

pressibility, C = 3 x

Near-wellbore damage or stimulation also makes the early part of the pressure buildup unreliable for determining formation pernieability. Con- sequently, the late part of the buildup, essentially the last 10 or 15 psi extending from point H of Fig. 8-9 and beyond, is critical. Recording must be continued until pressure returns t o within a few psi of the apparent reservoir pressure, as guessed by extrapolating the units scale of the digital recording. Otherwise, valuable information will be lost.

Permeability determined from the late part of the pressure buildup reflects primarily the flow taking place several feet out in the formation be>-ond normal invasion. The value so obtained should be the absolute permeability to hydrocarbon, as indicated in Fig. 8-1, if the undisturbed formation is at irreducible water qaturation. In this case the buildup permeability should be higher than the drawdown value if there is no local stimulation caused by the probe.

The method of calculating buildup permeability depends on whether the repressurizing flow is considered to be spherical or radial. The flow startsessentiallyspherical close to the probe. But asit moves outward it may encounter upper and lower impermeable beds that cause it to become radial


Fig. 8-15 Example of gas in flow line or pretest chambers (courtesy Schlum berger)

in the late stages. B> comparing radial arid spherical flou analyses, tlir nature of the late flow can bc deduced.

Radial Flow

given by

where '1 =

P Z

rn =

h =

Assuming radial-c~~liridrical flow, radial permeability, in riiillidarcies, is g!i$C5 gJod $U\W*J

!ir == -88.4 q . p/(tn ' h) (8.26)

average flow rate diiririg drawdown, i.e., total chamber volunic divided \>y total flow time, ccisec viscosity of the repressurizirig fluid, cp. It is the value for \vater at the temperature of interest if the forniation is water-bearing or the value for li!,drocarbon iri a ~iydrocarbori-beariIig formation nritli shallow in\-asiori. If invasion is deep, it should he a n iiiterme- diatc value slope of the tiormi. buildup plot, psiicycie, where pressure on a liiiear scale is plotted against the time function, (T -t At)/At, on a logarithmic scale. T is the total flowing time (24 see on Fig. 8-9) arid At is time elapsed since cessation of flow, i.e., time beginning at point G on Fig. 8-9. Pressure is read at At increments arid (T + At)/At vs pressure is tahilated iii preparation for making the Horner plot vertical thickness of formation sampled, ft. This is a variable figure since it is not defined by the tool but by the separation between tipper and lower impermeable beds straddling the test point. It could be as much as several feet in a homogenous forma-

t.,.,t't

PREDICTION O F PRODUClBlLlTY 295

tion and as little as ai1 inch in a thin sand bed confined b y shale stringers. In lieu of other information, it is generally taken as 0.5 ft

Table 8-1 tabulates pressure vs (T + At)/At for the example in Fig. 8-9 in At increments of 6 sec. Values for At greater than -18 see, which arecritical, are from the very late portion of the buildup not shown in Fig. 8-9.

The corresponding Horner plot is shown in Fig. 8-16 with time after flow progressing from left to right. A straight line is drawn through the late-

FORMATION P R E S S U R E = '\ 3857.1 P S I

3857.1 - 3843.0 5 LC) FE = Loglo 1 - Lcg10 2

/ 1 4 ' ' .. - 46.8 PSI = I__ -.. - - ,301 CYCLE

u,L 9 8 7 6 5 4 3 t + A t 2

- rn a

I

Y c4 2 u, w w L

t- Y =

3850

3840

3830

-AT-

Fig. 8-16 Radial flow pressure buildup plot of Homer (courtesy Schlumberger)

PREDICTION OF PRODUClBlLlTY 297 296 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION I

TABLE 8-1 TABULATION O F PRESSURES AND TIME FUNCTIONS FOR RADIAL AND SPHERICAL FLOW PLOTS

Radial (Horner) Spherical

D 1 + At 2 4 4 (psi)

3,824 3,834 3,839 3,843 3.845 3.846 3 847 3,848 3.849 3,850 3.851 3,851.3 3,851 7 3.852

- At 7 + At At 8 + At

6 30 500 14 12 36 300 20 18 42 233 26 24 48 200 32 30 54 18 38 34 60 167 44 42 66 157 50 48 72 150 56 54 78 144 62 60 84 140 68 66 90 136 74 72 96 133 80 78 102 131 86 a4 $08 129 92

24 +Af 30 36 42 46 54 60 66 72 78 84 90 96 102 108

O8165 02673 O 1826 O366 05773 O2236 O 1667 O 186 O4714 01961 01543 O 121 04082 O1768 01443 00871 03651 O 1622 O 1361 00668

03086 O 1414 0 1231 00441

02722 O 1270 O 1132 00320 02582 O 1213 O 1091 00278 O2462 O 1162 O 1054 O0246 02357 O1118 01021 00218 O2264 O1076 O0990 O0496

O 3333. O 1507 O 1291 00535

02887 o 1336 o 1178 00373

time points on the plot, ig above it because of afterfl this case the drawdown pressure shows no sign of g

ng early-time points that may fall below or nd near-wellbore damage or stimulation. In

into the pretest d disappear in a

1, that is, to essentially infinite recovery time, gives the shutin reservoir pressure. In this example it is 3,857 psi.

Determination of the slope of the straight line in psiícycIe, where cycle means a factor of 10 on the logarithmic time scale, is a little tricky if the line does not extend over one complete cycle. In that ease the slope i s best obtained by taking the difference in pressure between the points at ( T i At) /At equal to 1 and 2 and dividing by (log 1 - log 2), which is 0-0.30 as illustrated on Fig. 8-16. In the example, m is obtained as -46.8 psi:cycle. Consequently for Fig. 8-9

q = 20 ccI24 sec = 0.83 cclsec p = 1.0 cp, assumed h = 0.5 ft, assumed temporarily

Therefore, a first calculation of radial permeability from Eq. 8.26 gives

Ir, = (-88.4)(0.83)(1.0)/(-46.8)(0.5) = 3.2 md

This will be recalculated after a value of h is derived from the combined radial and spherical plots.

Spherical Flow Assiimingspherical flow during the late stages of repressurizing, spheri-

n)9'( (8.27)

cal permeability is given by the relation

where

q, ,u = as defined for radial flow m = slope of a buildup plot, p s i - v z , wherein pressure i5 plotted

linearly against a time function. f(t), given by - f(t) = (%/q,)/& - (%/ql - 1)/qT2 + At - 1/'-/Ti + T, +At (8.28)

in which qi and 92 are the first and second flow rates, T, and T, are the first and second flow times, and At i s the time after cessation of flow*

4 = porosity, fraction C=

For the example of ~ / q , = 2, T, = 16, andT, = 8so that

f(t) = 2 t d Z - l/iiS+iit - 1 , G (8.29)

Value5 of P vs f{t) are tabulated in Table 10-1 and are plotted in Fig. 8-17. The early points define a straight line extrapolating to a lower shutin

pressiire (3,851.0 psi) than indicated on the Horner plot, and the last points define a concave-upward curve. This behavior, combined with the late points defining a straight line on the Horner plot, signifies that the late flow is radial rather than spherical. If the late flow were spherical, the final points on the spherical plot would define a straight line, extrapolating to the proper shutin pressure, and the Horner plot would be a concave-downward curve extrapolating to a lower pressure.

The slope of thestraight-line portion of the spherical plot is the appropriate value form in Eq. 8.27. For this case m = - 104 psi-&, as indicated on Fig. 8-17. With the following values of other parameters

*For a single rate test. Tz = O and the equation ~implifit~r to

f(t) = l/& -l,JT+at where T is the flowing time and At the time after flow

298 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION PREDICTION O F PRODUClBlLlTY 299

4 = 0.10 [ L = 1.0 cp C = p i ' for a mixture of water and oil

Fig. 8-17 Spherical flow pressure buildup plot

I060

1850

- v) Q

Li <I)

?? a

3840

3830

1.~1. 6.27 gives

k, == (1,856)(1.0)~0.83/101)"(0.10 x 10~-'))" = 0.74 md

The spht,rical permeability, k,, is actuall!. a conibination of the radial per~neabilit!., k,, and \.ertical permeability, k,, in accordaiice with the re1 at ion

k, = (k,' , k,)'~3 (8.30 1

l ' ~ y i c d l > ~ ~ k, is significaiitl' lower thaii k, by a factor from 2 to 10 due to grain orienting on depositiori. M ' e shall assume an anisotropy factor. k,ík,, equal to 0.5.

Estimation of Formation Thickness Aii estimate can now be made of the thickness, h, of the formatiori

sampled. It is based on the difference in pressures obtained by extrapolating the straight lines on the Eiorrier and spherkal plots to At = 00 on the time axis. Assurniiig the probe is in the center of the bed, its thickness is given by

h = 0.039 {VA/[4n . (Pi, - P,). @ . C]}" (8.31)

\\..liere

V = volume of fluid taken into the tool, cc A = anisotropy ratio, k,ik, P,, = extrapolated pressure on the Horiier plot, psi í'< = extrapolated pressure o11 the spherical plot, psi Q = porosity C = compressibility of reservoir fluid, psi-'

Fortlieexample, V = 20, A = 0.5, P, = 3,857, P, = 3,851.0,a = 0.10, and C: = 10 .', which gives

11 = 1.9f t

tl~.calciilating the radial permeability with this value of h leads to

k, == 0.84 nid

For comparisori

k, = 0.75 md k ( d r a ~ v d o w n ) = 0.9 nid

All of these values are consistent, which gives some confidence that they are close to the true \,aliie.

With a gauge of 1 psi resolution, the maximurii permeability deterrnina- ble from pretest buildup is approximately 10 rnd (correspondirig to a ~nirii- mum measurable slope of 10 pdcycle on ttie iiorner plot). The range is


extended to 100 md with a gauge of O. 1 psi resolution. To reach beyond that limit requires that permeability be measured on sample buildup.

Advantages of Measuring Permeability After Sampling

1) Flow rates (q) can be 10-100 times greater into the sample chamber than into the pretest chambers, which extends the permeability range in direct proportion, as indicated by Eqs. 8.26 and 8.27. 2) Viscosities and compressibilities to be used in the permeability equations can be determined from analysis of the fluid sample at the surface. The viscosity of a mixed sample is

(8.32)

ic fractions of gas, oil, and water d pressure; pgt pQ, and p,, are their

(Vg - P$) + (VO . P”) 4- (V, ’ P w )

reconverted to rese viscosities at reserv

c, = (V, * cg + v, * C”) i- (V, * C,) (8.33)

where C,, C,, and C , are gas conditions. Charts are availa gas, oil, and water at reservoir temperatures and pressures.”

The quantity of gas in the sample can have a large effect on the parameters, decreasing mixture viscosity and increasing mixture compressibility. The sample may not be fully representative of the fluid investigated in the buildup, but it i s the best estimate that can be made. 3) The volume of rock investigated by the buildup is proportional to the quantity of fluid withdrawn, V(cc), which is far greater for sampling than for pretesting. The radius of investigation is given by

r, = 0.020 [V/(47r * 6p 4 * C]” (8.34)

where 6p is the gauge resolution in psi. With 6p = 1, the radius investigated in a formation of average porosity and compressibility (4 = 0.20, C = lo-’ psi-’) i s 2 ft for pretest buildup (V = 20 cc) and 15 ft for buildup following 23/,-gal sampling (10,000 cc). The latter radius extends well beyond any invasion effects.

It i s important when sampling to close the seal valve to the sample chamber purposely either before or at the time the pressure starts to build up from flowing pressure and to mark this time on the log as cessation of flow. Otherwise, afterflow obscures the start of buildup and may make much of

PREDICTION OF PRODUClBlLlTY 301

the buildup plot unreliable, especially if the chamber contains gas. The flow rate, q, to be used in the buildup equations is the total volume of fluid recovered, in cc, converted to reservoir conditions and divided by the total sampling time.

Tliere has been a good deal of skepticism concerning the reliability of pretest permeabilities because of the small amount of fluid withdrawn, the uncertainty in formation thick nipled, and the inaccuracy in buildup measurement. However. mat e improving. loois are being introduced with pressure resolution increased b!. a factor of 10 (O. 1 psi vs 1 psi) and with higher preteit flow rates and volumes. Horner and spherical plots can now be computer generated at the wellsite. Thickness can be estimated from the dual plots. All of these improvements should facilitate more accurate and rapid evaluation.

single measurement is insufficient to characterize the of a zone because permeability can vary drastically ne measurement every 2 ft in a hydrocarbon-bearing

zone of interest is desirable.

permeability from the Formation Tester and i porosity, into Fig. 8-Ga (or an equivalent) to d saturation, S,,,.“ This. along with actual S,, from logs, is then entered into charts such a$ Fig. 8-2 to predict water cut. The method is appealing. brit the inaccuracies are we11 that i t ran onl> be successful i f careftill- caiibratcd against pi ior p r o d u tiori in a qiven fieici.

A technique that has been used for productivity prediction is toco

SAMPLING AND SAMPLE ANALYSIS Fluid sampling M ith the RFT is aceoniplislicd 1)) opening oncb oi the s r d

\alve\ (Fig. 8-8) and ailo~ring reservoir tiiiid to flom into tlie selectctl chan;ber.”As the fluidenter5 the upper part of thechamber, it forces a float- ing piston to expel water from tlie lower part, through a choke or seleral chokes in series, and into a ballast chamber below that is initially filled with air at atmospheric pressure. The size of the choke is chosen to Iestrict formation flow rate. if neccísary, to prel ent breakdown of soft formations with high permeabilities.

Prior to wnpling, an eitimate of the flow rate and fiov ing pressuie t o tw expt3ctd can be obtained f iom Fig. 8-18 Shutin rewrvoir pressurc. P,. is marked on the iertical scale, and a x aliie <in,&, i s marked on the horizontal scale \T. here

(3

O o

cu O

m

x

P

n

VI

n

3

P

-

3

3

3

N

3

- 3

304 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION i 1

attached. This consists of a pressure regulator, a calibrated plastic bottle, and a gas meter. When the valve is opened, the fluid sample is forced out of the chamber through the separator. Liquid drops out in the bottle and the

water on bottom and oil on top, and the amount of each is read from a graduated scale on the side of the bottle in cc.

Calculating GOR and Water Cut

From therecovered amounts the gas-oil ratio (GOR) in cu ftibhl and the water cut (WC) can be computed. These are given by

COR = 159,000 x (cu ft of gas)/(cc of oil) (8.36)

(8.37) cc of formation water cc of formation water + cc of oil we =

ontain a substantial amount of mud les were withdrawn at the same trate would drop to zero but the

would remain the same. of mud filtrate must be r recovery to obtain the

quantity of formation water. To determine the fraction of formation water, the logging engineer

measures the resistivity, Rri. of the recovered water in a resistivity cell. This is compared to the known resistivities of mud filtrate, Rmf, and formation water, R,, all values being converted to the same temperature. If Rri = Rmr, the water is all filtrate; if Rrf = R,, it is all formation water. For the usual intermediate case, the ratios RmI/RrI and RJR, are entered on the two axes of Fig. 8-20 and the percentageof formation water i s read from the numbers on the curved lines. As an example, assume the following recovery

Gas: 0.51 cuf t Oil: 350cc Water: 3,200 cc

Resistivities

Rrf = 0.54 ohm-m at 75°F Rmf = 0.76 ohm-m at 75°F R, = 0.080 ohm-m at 200°F (formation temperature)

PREDICTION OF PRODUClBlLlTY 305

The gas-oil ratio is directly

O 51 350 COR = x 150,000 = 230 cu f thb i

R , = O.OS0 (ZOO + 6.7)/(75 + 6.7) = 0.20

R, - R, - R,, must be at carne temperature

2 2 5 3 4 6 8 IOR,,, 20 40 60 80 ___) -

R.

Fig 8-20 Determination of formation water percentage in recovered sample (courtesy Schiumberger)

306

Ratios arc'

prediction of hydrocarbon production with small hydrocarbon recoveries is reliable only with a good deal of local experience.

ESSENTIALS OF M O D E R N OPEN-HOLE LOG INTERPRETATION

Repeatability: _t 5 psi

PREDICTION OF PRODUCIBILI'I'Y 307

A P P L I C A T I O N S O F PRESSURE MEASUREMENTS

11 ,,,, Ii,, = 0.7w0.54 = 1 .il R . , , , R . = O.TCjíO.20 = 3.8

Perhaps the most important feature of the formation tester is its ability to measure reservoir pressures at multiple levels much more quickly and less

308 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION PREDICTION OF PRODUCiBlLlTY 309

Absolute accuracy: +25 psi with normal field calibration k 15 psi with special field calibration

As already illustrated, a 1-psi resolution is barely adequate for

permeability determination. Pressure gradients in permeable zones of a given well vary from approxi-

mately 0.46 psilft for salt water to 0.40 psilft for light oil to 0.10 psiÍft for gas. Over a 20-ft interval the pressure differential from top to bottom would be 9.2 psi for water, 8 psi for oil, and 2 psi for gas. Distinguishing between water and oil i s therefore difficult with the standard strain gauge of 1-psi resolution. It really demands a gauge of O. 1-psi resolution and also demands that thestability of the gauge under these circumstances be far better than 5

producing reservoirs this may cause no problem since pressure differentials may be hundreds or thousands of psi apart in correlatable intervals that do not communicate. In virgin reservoirs: however, much better absolute accuracy is required.

The High-Accuracy Gauge

For high-accuracy applications a quartz-crystal gauge (also called the Hewlett-Packard or HP gauge) can be run with the RFT. Its characteristics are resolution of f0 .01 psi, repeatability of k0.4 psi, and absolute accuracy CJf rt0.5 psi.

This gauge has amazing precision-10 to 100 times better than that of a strain gauge. It would be in universal use except for two factors. One is that it is expensive and fragile. The other is that it has a settling time of about 10 min. That is, each time a pressure change occurs, the gauge requires 10 min or so to stabilize. This makes it useless for short-term pressure buildup measurements. However, it is ideal for intrawell pressure gradient or interwell absolute pressure recording. The prime use of this gauge, however (not necessarily in the RFT), is for interference tests whcre a flow rate

change is made in a given well and the corresponding pressure change is observed in an adjacent well. Such tests, which require extreme resolution and stability in pressure measurement, are very valuable in establishing

Single Well Pressure Applications

Ozjerpressiiritig: Formations ma)' be highly overpressured to the extent that pore pressure may be twice the hydrostatic head of salt water at that depth. While this must be anticipated prior to drilling, once a wildcat is drilled it i s desirable to obtain a pressure profile in the well to optimize both production and offset drilling. For this purpose great accuracy is not required. The strain gauge is adequate.

gauge. In measuring pressure gradients, several points must be kept in mind.

First, the gradient of the mobile fluid phase behind the in\ aded zone is beins measured. The irreducible \vater does not influence the reading because i t i> held in place by surface tension, nor does the presence of nitid filtrate influence the pressure-the invaded zone is too thin. Second. in a low- permeability zone reservoir pressure must be obtained by extrapolation of the Horner or spherical plot to obtain an accurate t aliie. Third, i n extremely low permeabilities the extrapolated pressure can be abnormally high because of supercharging.

Supercharging is temporary overpressuring caused by invasion. The differential pressure between mud and reservoir at a given levei. which is reyponsible for invasion. is divided between the mud cake and the forination according to their respective flow resistances. Normally the mud cake has much higher resistance than the formation so that the excess pressure created in the latter is negligible. However, once formation permeability drops below about 1 md, this condition no longer holds and the formation next to the borehole may become supercharged. The magnitude is strongly dependent on downhole water-loss characterisitics of the mud and is not


1iredictni)le. Ho\vever. it is bclie\,cd to tie on thr oi-der of l / k p i wlieick i s the formation pernir,ubility in nid.

Point 1-1 of Fig. 8-21 is a ease in point. I'ernieability at that le\rl \vas foiiiid to he 0.024 nid, \5 liicii w-oiild lead to an estimate of -10 psi uf supercharging. Poiiits siich as these iriiist he ignored on a gradient plot.

Prc.v.srrrr drpk t io i i : I n a rttsrr\,oir prodricing from sc\wal zones, IJres- sure rnrasiireriierits in infil l or olxcrvatiori \veils providt. a n opportiiiiity to inonitor the depletion of the individual intervals. Fig. 8-22 sho\vs a case \\,here zones depleted at quite different rates. Sudi information is needed to plan stimulation of nonprodiicirig zones or to prevent depleted intcmxls froin becoming thief zones.

- al al c

n > k-

7500

7600

7700

7800

'15 \ x - + @ l 4

\ -

\ x 7 @ Extrapolated pressure

x 6

pn 0 2 9 p s i f t - O 67 g cc I-\ ' x 5

Q x */7 'x 4

\ Extrapolated for temp stabilization

\

( I R --

047 1 o9 g cc I-\ Not temp. stabilized 4 ~

\

3550 3600 3650 3700 Formation pressure - psi

Fig. 8-21 Pressure gradients in a Middle Eust reservoir (courtesy Sctilumberger, O SPE-AIME)

PREDICTION O F PRODUCIBILITY 31 1

3000

5 3100 Q) ..- E - r_ a n

3200

3300

\ \

l U Y \

\

\ \ - \ l 3 7 Carbonates

\ \

\

\ Sandstone

\ \ \ 8'L Carbonate

\ 7 A Sandstones

. \ ' Carbonate \ \

\ "1

<andstones - 'L I I 1-L-I_-

4500 4600 4700 4800 4900 5000 5100 Pressure (psi)

Fig 8-22 Pressures in various producing zones of a Middle East well after severui years of production (courtesy Sctiiumberger, (3 SPE-AIME)

Multiple Well Pressure Applications

Idt~ally Ivhen exploration ~ v e l l s are clrilled to define reservoir limits, pressiire ineasiirernents should he made t o determine \vlietlier the prodiic- t i w liorizons defined b y the logs actually corrinitinicate between \veils. Accurate rneaswenients of static pressiires are good; interference tests are even lwtter. This practice is slowly advancing.

i t 1x.coiries imperative to know interwell coriimuiiication \$.hen secondary recovcry, usually waterfloding, is undertaken. hlany cases are documented where unsuspected permeability blocks caused poor perforin- anee of the flood. Fig. 8-23 illustrates pressure recordings over a 600-ft produciiig interval in 12 infill wells drilled into ai1 old reservoir in Colorado preparatory to additional water flooding. The separation between vertical lines represents 4,000 psi. Great lateral as well as vertical variations in

312 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION PREDICTION OF PRODUCIBILITY 313

I I 1 w 3 i

i I I

Fig. 8-23 Pressure profiles across a producing field, Colorado (courtesy Schlumberger. O SPE-AIME)

pressure are observed, indicative of many permeability barriers. For example, the lowcrmost sand in Zone B differs in pressure by about 2,000 psi between adjacent wells 11 and 12, even though they are at the same strati-

Forthcoming Toot Improvements Other applications of formation tester pressure measurements are

continually coming to light.26 Each advancement in tool accuracy and versatility opens new avenues. Major improvements in the RFT just being introduced are

1. Pressure resolution of O. 1 p s i and absolute accuracy of _+ 10 psi (with

2. Fast flow option with a probe approximately five times larger in area monthly deadweight tester calibration).

e flow rates accor d the measurable

of 10 or 20 cc. Thi ncrease the radius

'

ary ga-

(Eq. 8.34). However, as the pretest chamber

These and other mechanical improvements should significantly improve the usefulness of the HFT and similar tools.

is exhausted into the sample chamber after each pretest.

SUMMARY

The Multiple Formation Tester is a tool that can be run in open hoie, positioned at a desired level, and actuated to take a sample of reservoir fluid f2-12 gal] and to measure reservoir fluid pressure.

Only two fluid samples can be brought back to the surface for analysis. However, any number of pressure measurements can be made at any level desired. Information obtainable is

1) From analysis of fluid samples at the surface An estimate of the gas-oil ratio expected on production An estimate of the water cut expected on production

314 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION PREDICTION OF PRODUClBiLlTY 31 5

The sample normally contains a substantial amount of mud filtrate, The fraction of recovered water that is virgin water is obtained b y comparing the resistivity of the recovered sample, R,f, with that of the mud filtrate, R,,, and the formation water, R , .

2) From downhole pressure measurements Reservoi r permeabi I ity Reservoir pressure Location of gas-oil-water contacts Communication between zones

The MFT is the best wire-line method of obtaining permeability. It also gives valuable information on reservoir connectivity for secondary recovery.

REFERENCES

‘T .C. Frick, Petroíc,irtri I’rodiidioti lforidbook (New York: XlcGra\v-IIill, lOG2), chapter 23.

‘F , rick, chapter 25, Ibid. ‘J. Pelissier-<:oiribt.scure, D. Pollock, and X I . \\’ittinan, “Application of

Repeat Forriiation Tester Pressure hleastireirients iii the hfiddle East,” SPE 7775, hlanaina Bahrain (March 1979).

‘J.P. Jnnes, “Production Engineering arid Reservoir hlecliariics,” Oí:]

’H.1,. hlorris alid \T’.P. Riggs, “Ikirig I.oq-Ikrived Valiit.5 of \l’ater Satiira-

‘D. K . Keelari, “Core Analysis for Aid i n íkaervoir I>t.scription,”]. Pot. 7’wI~.

‘Frick, ctiapter 23, Ibid. ’1l.P. hfurpli). a r i d \\’.\V. O w < m , “A N e w Approach for Lon- Resisti\*ity Saiitl

‘htorris and Biggs, lbid.

(1915), pp. 45-46.

tion and Porosity,” Sí’\I’í,/\ Logging S y t r i p n i t r r r i Srunsactiotis (i96;)

(November 1982), pp. 3483--2401.

Log Analysis,” S P E 3560, Ne\\, Orleans (October 1971 ) .

A. Tiriiiir, “An Investigation of Prrineabilít>, Porosity and Residual \\’;iter Saturation Relatiorithips,” Sl)\VLA Loggirig S y n t ~ ~ ~ i s i i u r r i Srotiwcfioris (Jiine 1968).

Scliluni1)erger. Log 1ritr.rprettrfiuti Charts (1979), p. 83. ‘ k . R . Cnates and J .L. Durnanoir, “A New Approacli to Iinpioved Log-

Derived Permeability,’’ SPWLA Logging Syr t iywsiu tn Transoctioris (&lay 1973). I3L.I,. Raymer, “Elevation and Hydrocarbon Density Correction for Log-

Ilerived Permeability Relationships,” The Log A7iolyst (May--June 1981), pp.3-7. ‘*A. Brown and S. Hiisseini, “Perineability from Well Logs, Shaybati Field,

Saudi Arabia,” S P W L A Logging Symposium Trunsuctioiis (Jiine 1977).

I o

11

J . J . Sirioleri anJ I . . R . L,itsry, “Formation Evaliiatioii Using M‘irelirie 1;or- iiiation Tester Pressure Data,” SI’E GSO2 (Denver: October 1977).

D.K. Sethi, LV,C. L‘ercellino, arid \ \ ~ . I I . l;ertl, “The Formation híulti- Tc.stt,r -- Its Basic Principles arid Prx t ica l Field Applications.” CPLVLA Loggi r ig S!! t t i p i t i t i l 7‘rarisartiori.s (1 111;; 1980)

J . € i t Lloran aiid E . E , Finklea, “Theoretical Analysis of Pressure Phenomena Associated witti the \Vireline Formation ‘I’cster,” J . Pet. Sewh. (Augiist 1962). pp.

15

i o

1;

1-908. l’elissier-Coriibesciire et al., Ihid. G . Ste\vart arid M. \Vittrnaiin, “Interpretation of the Pressure Reiponse of

I h

19 7

tiie Repcat Formation Tester,” Si’E 8362 (Las Vegas: September 1979). %‘rick, lbid. ”C . \ V . I.,ock\vood and D.E. Cüri i ion, “Production Forecastiiig,” SPF: %)*ti,

Bakersfield (March 1981). --Scliluniherger, “Formation ‘Tester Interpretation Methods and Charts,”

paper C-11721, ‘“Snioien and Iitsey, Ibici. LIPeIissier-Conit,esciire et al . , ItjicI. ”í;. Stewart arid L. Abwtaraii, me íriterpretatiori of Vertical Presm-e

Cradients Measiired at Ol)st.r\,atioii Wttlls in lkveloped Reservoirs,” SPE 11131, Ne\v Orleans (September 1982).

“C. Stewart, J. R’ittmann, and T. \’an Golf-Racht, “The Application of the Repeat Formation Tcster to tlie Analysis of Naturall!. Fractured Reservoirs,” SPE 10181 (Sari Antonio: October 1981).

“ 9

ellsite computed logs are often called “quick-look” logs. They are curves designed to show directly the presence of hydrocarbons. Their princirial use is to pick intervals for testing or sidewaIl u

sampling and to facilitate a quick decision on whether to set pipe and complete the well or to plug and abandon it.

Before wellsite cornpilters there were four types of quick-look logs in use

the apparent water resistivity, R,,,, log

the SP or R,,,íR, overlay

log computing circuits vasion, or shale effects. t was made to correct lo

The introdiiction of digital computers to surface logging units has brought much greater versatility to the generation of wellsite-computed wrces. The basic logs are recorded digitally on tape as they are tieing run. Following the logging runs. they can be played back, edited, corrected, merged, and coinputed, and outputs can be presented in a variety of for- mats. Consequently. there has been a veritable explosion in the variety of wellsite-computed curlyes in recent years. Programs are available for shaly formation interpretation, true vertical depth adjustment, matrix identification, interpretation of formation tests and dip conputation, to name only the more important categories,‘ ‘A fully interpreted log can be acailahle to the \vel1 operator two hours after logging is compfeted.

The four quick-look logs listed above are described in this section, follo\ved by an outline of the general-purpose welisite interpretation pro- gramy. The Schliiniberger version of the latter is called CJ*berlook and the Dresser-Atlas version is called Prolog.

31 7

31 8 ESSENTIALS O F M O D E R N OPEN-HOLE L O G INTERPRETATION WELLSITE C O M P U T E D L O G S 319

THE R,, LOG

The R,, log is a quick-scan ciirve used for direct indication of hydrocarbon-bearing zones, determination of water resistivity, R,", and estimation of water saturation, S, R,,, stands for apparent water resistivity. The R,, curve is obtained by continuously calculating H,, on the assumption that S,v = 1 everyLvhere. The Archie equation is

Placing S, = 1, replacing R, by R,\,, and solving for the latter gives

E,,, = (4 . r\,)/c? (9.2)

Typically, the R,,, curve is recorded when either Sonic or Drnsity-Neu- tron logs are run siniultaneously with the Iiidiiction log. Porosity is obtained from the porosity logs and R, is obtained froni the deep Induction log. R,, is continuously calculated using Eq. 9.2. It is usually presented in Track 1.

If all permeable formations are clean and water bearing, the R,, log is a slowly varying curve whoqe values are tlie actual K,, values in those formations. Thus, R,$ in a given interval can usually be read directly from the R,, curve.

If , however, a hydrocarbon-bearing zone is penetrated, R, will be greater than it would be if the zone were water bearing and the R,, value derived from Eq. 9.2 will be higher than R,. Water saturation can be directly ebtiinated from the R,,,/R,, ratio Eq. 9.1 may be rewritten

Replacing (4 ?IR,) by its value as given b y Eq. 9.2, namely c'R,,, gives

Therefore

(9.4)

If R,,, = R,, then S,, = LOO'? = 2 I{,$ == 70 % = 3 I{, = 58"ó = 4 R, = 50 '/o

Consequently, the R,, curve can be scaled directly in water saturation, at least over short intervals (a few Iiundred feet), where R, i s constant. Alternatively, a rule of thumb is that if R,, > 3R,, comniercial production is possible and a thorough analysis of the zone should be made.

Fig. $1- 1 h \ v s a11 R,,, curve recortlrcl \vith tlie ISF-Sonic co~nbiiilitiori. IR,\ ~ v o d d be read as 0.05 ohm-ni in the \vater-bearing zone ( 2 ) . The critical I{,,,, \.alrie is therefore 0. 15 o h - m . The ~iydrocarbon-bearirigsection clearly sho\?,s at levels 7 and 6. A t level 7 , It,, = 0.5 so that s,, = t 'O .05 '0 .5 or 0.32. The thin zones above and below those levels, Lvith R,,, > 0.15, are not hydrocarbon hearing. They result from the fact that the Sonic responds to tliiiiner zones than the Induction (2.5 f t vs 5 f t ) . Averaging the Sonic over 5 f t would eliminate these spurious indications.

When the Sonic is the porosity tool, Schliimberger uses the relation given by Eq. 5.18 to convert travel tinie to porosity. This avoids the need for a compaction correction in uiiconsolidated sands. \$'hen the Density-Neutron combination is used for porosity, the arithmetic average of tlie and 4, values is used. In both cases a certain amount of shaliness is tolerable since both Sonic and Dtmsity-Neutron tools will give porosities somewhat high but the shaliness will generally cause R, to be too low. The effects tend to compensate, and Fiw, R,,,, and S, values are reasonably good. This is not true if Density only is used for porosity. S,. values will be abnorrnally high.

The R,, curve works \vel1 in rnedium to high porosity zones where matrix type is fairly constant (Gulf Coast). It is not appropriate for low-porosity regions \\,here matrix is quite variable because porosity values will be too matrix dependent. This is especiall>r true if Sonic or Density is used for porosity, lcss so if Density-Neutron is used.

POROSITY OVERLAYS

A porosity overlay is a presentation of two porosity curves on the same track with the same scale. Presentation will bc in Track 3 if the curves are run siniultarieously with resistivity or in Tracks 2 and 3 if they art: run st-parately.

Porosity overlays provide a quick visual indication of matrix, shale, and hydrocarbon effects as well as porosity. The three possible porosity combi- iiations are Density-Neutron, Density-Sonic, and Sonic-Neutron. Of these the Density-Neutron overlay is the most useful for several reasons. First, the true porosity is very close to the average of Density and Neutron readings, regardless of the lithology (shaliness and anhydrite or gypsum excluded). Second, lithology is often evident from the relative positions of the two curves. Third, gas-bearing zones stand out.

Fig. 9-2 illustrates the first two points. The beds at 14,553, 14,564, 13,650 ft and similar z o n a are clearly anhydrites with zero porosity. The levels at 14,504 and 14,680 ft are dolomites (somewhat radioactive, as

320 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION WELLSITE C O M P U T E D LOGS 321

I 10 .

Delaii log 5 , " - 1OOn

6400,

6 -

5 4

4 4

3 -

2 -

> .

6930

1 1 5 0 I so

Fig!9-1 A hydrocarbon zone indicated b y the R,, curve (courtesy Schlumberger)

15 ------ 7- -- --- -r------------ Diam in inches

r-r-T-r-rT-rvv

14700

'Fig. 9-2 Lithology changes indicated b y a Density-Neutron porosity overlay (courtesy Schlumberger)

P, v

3 iz

-8

I 1

N h . v U

a

’ WELLSITE C O M P U T E D L O G S 325 324 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION

Thus, R, is generated continuously by obtaining $J from the porosity log and combining it with R,, as indicated.

A comparison of calculated R, in a given zone with measured R, immediately gives water saturation

s,\ = t/R,R, (9.6)

If R, = R, then S,, = 100%a = 2 R , = 70% = 3 R,, = 58% = 4 R, = 5070

Commercial production is possible if R, > 3R,). On the logarithmic scale used for resistivity, a given RJR, ratio represents a constant separation

F = (e/&)’ (9.7)

By definition, F= R,íR,. Consequently, when the calculated F is recorded on the logarithmic resistivity scale and is shifted to match the measured R, curve in water-bearing zones (which essentially multiplies F by R,,). the curve becomes an R, log. Separation between R, and R, shows hydrocarbon- bearing zones as described above.

As with R,, curves, resistivity overlays recorded with thcx Density log as the source of porosity have no automatic compensation for Fhaliness. Water saturations in shaly sands tend to be pessimistic. Values i f 4 1 be more correct with Density-Neutron or Sonic porosities.

THE SP OVERLAY

‘The SP overlay, also called the RJR, or SP quick-look, is a comparison of a computed SP curve with the recorded SP log such that departures directly indicate the presence of movable hydrocarbon^.^ It is best used in conjunction with R,, or resistivity overlays, which respond to total hydrocarbons.

A

B

Fig. 9-4 Hydrocarbon presence indicated b y resistivity overlay and movability by SP overlay (courtesy Schlumberger and SPWLA)


hlovable 1i)di-ocarboris are indicated by a difference between iiridis- tiirbed-zone \vater saturation, S,, , aiid flushed-zone saturation, S,,,. If S,,/S,,, = 1, the zone \\,ill prodiice water, regardless of apparent water saturation. If S,,/S,,, < 1, the zone contains movable hydrocarbons. The comparison is accwnplished as follo\vs

S,, = c J.,i.,b (9.8)

Dividing both sides arid rearranging

(9.9)

S,/S,” = [(RiJR,)/ml,~l mJ1’ (9. 10)

The logarithm of the ratio R,,,,/R, is routinely presented in Track 1 as the SI’ since

SP = - K lüg(R,,,JR,) (9.11)

The object is to compare log R,,/R, with the SP. T>pically, the Inductiori tool is run and the ratio RJR, i\ obtained from the shallow resistivity reading (16”N, LL8, or SFL) and the deep resistivity reading (ILJ. A reasonable value of RJR, can be derived, provided the invasion diameter is between 20 and 100 in. even though the shallow curves read beyond the flushed zone. An on-site, continuous computation is made to give a quick-look SI’

SI’, = - K log(R,,,/R,) (9.12)

The SI’, curve, compatibly healed, is recorded in Track 1 along with the normal SP.

In a water zone where S,/S,,, = 1, the ratios RJR, and R,, JR, are equal and the curies converge. In a zone mith movable hydrocarbons, S,,IS,,, < 1. Thus, RJR, becomes less than R,,,,/R, and the curves diverge, with the quick-look SP,, becoming smaller than the recorded SP.

Fig. 9-4 includes the SP quick-look curve. The separation between that curve and the SP in Sand A indicates the hydrocarbons can move.

The SP overlay has several advantages relative to R,, and resistivity quick-looks. It indicates movable hydrocarbons, not just the presence of hydrocarbons. It is independent of porosity and lithology; therefore, it is more suited to variable-lithology formations than R,, or resistivity overlays. Chaliness is automatically compensated for since it affects both SP and SP, in much the same way. Finally, variations in R, are automatically taken into account.

WELLSITE C O M P U T E D L O G S 327

In spite of these aclvant ages, the SP quick-look has not found widespread ;icwpt:incc. Conit drawbacks are that the rncthod is limited to fresh muds \\.here R,,,, > I(,,, that the computation is sensitive to :il)nornia! invasion tfiamcters ( < 20 in. or > 100 i n . ) . arid tlint numerical valiies of S,,!S,,, cannot readily be estimated.

THE CYBERLOOK LOG

The Cyberlook log is a Schlumberger quick-look slialy formation analysis.5 It is based on theDual-Water niodel described in Chapter 7. Theoutput consists of a set of computed curves, the most important of which are shale- corrected water saturation and porosity. The program is suitable for both sand and carbonate anal‘

The basic nieasurenients utilized are Dual Iriduction-SFL (or an equivalent set), Density, Neutron, SP and CR. Optional nieasurenients are Sonic, which is desirable in rough hole situations: and R,<,. which allo\vs rno\rable oil computation.

Processing is carried out after the logging runs are completed. First, a Pass One set of curves is optically recorded. The purpose of this is to let the logging engineer select parameters needed for the final Pass Two interpretation run. For Pass One he needs only to specify the following constants: matrix to be used for computing Neutron and Density porosities (sand or lime), miid weight, bit size, pore fluid density, bottom-hole teniperature, SI’ drift, and depth offsets for merging.

In I’ass One the basic nieasurements are nierged. if they are obtained on different runs, and environniental corrections are applied. The Camma R a g . is correctcd for hole size and mud Lveight effect, the deep Induction is corrected for invasion. tlie Neutron is adjusted for teniperature and pressure effects, arid the Density is corrected for borehole size variations.

Fig. 9--5 shows portions of ISF-Sonic and CNL-FDC logs recorded over a n interval of Miocenesarids and shales in a South Louisiana well. Fig. 9-6a is the Pass One output for these logs. In Track 3 are the corrected Neiitron and Density porosity curves and an apparent total porosity ciirve, +,,, computed from tlie Neutron-Density crossplot. Track 1 contains the SI’ and corrected GR logs and an apparent grain density, pq“, curve computed from +,, and bulk density. In Track 2 is the corrected deep Induction log and an apparent fluid resistivity log, Rr,, computed from corrected €3, and &. The scales are quite different for the Rt and R,, curves; it just happens that the two curves overlap in this case.

I


hlaxiniiirn porosity = 36‘ I R,,, = 0.018 ohm-ni (clean wet sections of I: and C:) H,,,, = 0.09 ohm-ni (interval D)

Iii this case, parameter picking is fairly straight-forward. However, in sitiiations where there are no clean formations, where 11, is \,arying rapidly, or where niatrix is not constant, selection is very judgmental. In such instances the experience of the logging engineer in the region of interest is an import ant factor.

iI’ith the selected paranieters entered into the computer, the Pass Two run is made arid the final interpretation log is produced. Fig. 9-6b is the Cyberlook for the example case. l r i Track 1 is the shale index, essentially the lowest of the values computed from the G R and SP (and other, if used) shale indicators. 111 tlie same track is the computed grain density. This curve reads correctly in clean formations, but the indicated grain densities in shaly formatioris are abnornially high because the Xeutrori porosity includes the lattice-bound hydrogen in clay.

In Track 2 is a four-decade logarithniic resistivity overlay. The corrected deep Induction resistivity, R,, is reproduced as the dashed curve. Overlain on it, as a solid curve, is the wet resistivity, R,, coiiipiited from total porosity, bound-water fraction, arid free- and bound-water resistivities. This curve should overla!, the R, curve in all water-bearing zones but sliould show lolver resistivity values in hydrocar1)on-bearing zones. The latter are clearly shown in zones A , C , and F a s scparations between the R,, and R, curves. If the Ii!.drocarbon is gas, it is so indicated by a b a r in the depth column, as appears opposite the iipper sand of Zone F.

\i’atcxr saturation is recorded in the left portion of ‘Track 3. This is calciilüted itsing a simplified version of the Dual-\Vater eqciatioiis presented in Chapter 7. \VIit~n an R,,, log is run. the flushed zone watt can also be recorded alongside S,\; tlic difference between the cur\w indicates ino\.ahlr oil.

In the middle of Track 3 is a differential caliper ciirve that represents the difference br.tnwn itctiial hole size and bit size. The zero is in the center of the track.

On the right-hand side of Track 3 is the porosity information. The outer envelope of the ciirve toward the depth track represents effective porosity; the shaded section represents that portion of bulk volume which contains hydrocarbons; and the white section represents that portion containing water.

WELLSITE C O M P U T E D L O G S 331

The quicSk-looL appeal of tlie coriiputed log is quite evident, n-itli tlie Ii!,cIrocarboii-bearirig zones staridirig out in ldack. Iri this cast’ a11 of these intervals are indicated us good pay zones on the R,,, ciir\’e of tlie basic ISF- Sonic log. II«\vever, in shalier. lower-porosity formations such as t l i r h South Texas \Z‘iIcox, tlie quaiititatii~c information i n the computed log is rieedetl for completion decision.

Questions have been raised about deterinining bound-water fraction and rquations that are used for water saturation. Ne\wtheless, the aiis\vers agree fairl!. \$.ell wit11 those from more sophisticated p rogra~ns .~

THE PROLOG ANALYSIS

Prolog ih the Dresser-Atlas computerized wellsite log analysis s!.stein.’ I t coiisists of two separate packages: Wellsite SAND for clastic sedimeiit analysis arid li’ellsite CRA for carbonate sediment sequences.

The input to I’rolog and the final output are similar to those for Cyberlook, and the same operations of data merging and environniental log correction are applied. Hou,ever, in tlie l’rolog sjxteni a first pass or preinterpretatiori optical log is not made. Instead, t h e engineer selects a zone to be interpreted and proceeds through a dialogue with the computer i n Lvhich he specifies the necessary clean and shale \.slues for shale index determination, the matrix \dues for porosity determination, and the K,, value for [rater saturation calcrilatioii. The computer then calculates the desired outpiit, including shale fraction, cffective porosity, and watrr saturation, ant1 displays the result!; on the \,ideo monitor. If the engineer questions the computed \dues , h e readjusts inpiit parameters and repeats the calculation. Only .ivlieri he i‘ satisfied with the result is the computed data corrirriittcd to tape to beCoIrie part of the final output. This procedure is curried out for each interval into which tlie \vel1 has been zoned.

To help the engineer zone the well and choose correct values of R,, and matrix constants, the Prolog system allows the generation of various t!.pes of crossplots at the lvellsite, including bulk density 1’s Neutron porosity and porosity 1.s resistivity. For watrr saturation calculation, Wellsite SAND uses the Simandoux shaly-sand equation and Wellsite CRA uses the Fertl shaly for 111 at ioi i re1 at ion.

Fig. 9-7 shows the Wellsite SAND output. It is a four-track format. In Track 1 are SP and shale fraction. Also shown is integrated borehole volume (VBH) as pips, each one representing 10 cu f t .

332 ESSENTIALS OF MODERN OPEN-HOLE L O G INTERPRETATION 333 1 WELLSITE COMPUTED LOGS

Li<hologic characteristics

S W Bulk volume

U

SP

Depl - Porosity and lluids Formation

Water I 4 Bulk volume I O'O Bulk volume I

, ,O

HydrCCo tor leet

Dai r -44,

Fig. 9-7 Example of Wellsite SAND (courtesy Dresser, O SPE-AIME)

Ir1 the depth track integrated porosity feet (PF) is shown as pips on the left, and integrated hydrocarbon feet (HF) is shown as pips on the right, each representing 1 ft.

h enlargement.

Track 3 shows effective porosity (POR) and its breakdown into water (PORW) and hydrocarbon (blackened). Between Tracks 2 and 3 is the differential caliper (CALC) .

Track 4 is the bulk volume analysis, showing the porosity, sand, and clay components.

The Wellsite CRA output format is very sirnilar. A GR curve is included in Track 1. If an R,,,curve i s run, Track3 alsoshows theportion of porespace filled with mud filtrat

OFF

Wellsite-computed logc such as Cyberlook and Prolog are excellent for the purpose intended but are not rneant to give final answers. Where iriter- pretation is difficult, the best accuracy is desired, and the cost is justified, more coniplete log processing should be carried out at log interpretatíon centers. The major oil and Service companies llave such centers in s t ra tqie locations around the Lvorld. There, the processing is siipervised by exper- ienced analysts, interpretation algorithms are more sophisticated, and tnuch more time is allotted to generating histograms, crossplots and intermediate playbacks for zoning and parameter selcction.

For reference. office-based open-hole interpretation programs that are used by major oil or service companies and that have been described in the literature are listed be lo^. In each case the water saturation equation used is indicated.

Shell:

Scliliimberger:

Lamisi (for laminated or dispersed sand-shale: Waxman-

Saraband (for shaiy sands: modified Simandous rela-

Coriband (for shaly carbonates; Simandoux relation)" ''

Smits relation)'

tion)' lo

334 ESSENTIALS OF M O D E R N OPEN-HOLE LOG INTERPRETATION WELWTE C O M P U T E D L O G S 335

i’olaii (rq,!iicing tlic. previous t\ro fur sha1.i. sunrls or carboiiates: Duai-iVater equation)’

Global (an adapti\,e niinimuni-error program for stid!. sands or carhnates: choice of i$’uxinan-Siiiits or Dual- if’ater)’4

Dresser-Atlas: Epilog Saiid (for sliai!, santls: Siriiandoux relation)” Epilog CRA (for slialy carbonates: F‘crtl relation)’,’ Digital Shaly Sand Ai-ialysis (ii’axiiiaii-Srnits)’”

Refrrcsiice should he mnde to the indicated papers for the details of tlwsc. programs: they are quite complex. iiotvever, the f ina l conipiited logs are similar in presentation to those of C!.berl«ok or Prolog.

SUMMARY The following quick-look curves provide wellsite direct indication of hydrocarbon presence.

I . Apparent Water Resistivity, R,,, Requires deep resistivity and porosity input Indicates presence of hydrocarbons Works best in medium-soft rock, constant lithology Compensation for shaliness using or & :,, not d8> Direct reading of R,,, Commercial production possible if R,, > 3 R,

2 , Porosity Overlay Requires two porosity curves; Density-Neutron is best Indicates presence of gas b y crossover Shale and dolomite suppress crossover, may ObSCLJre gas False indication of gas if matrix is sandstone but recording is

Lithology variations evident in tight carbonates I i mestorie

3. Resistivity Overlay, R, vs R, Requires porosity and R, input Indicates presence of hydrocarbons Works best in medium-soft rock, constant lithology

Compensation for shaliness using & ri, not iD Cornn-ierciat production possible if R, > 3 R,

4 SP Overlay IRequires only deep and shallow resistivity input Indicates presence of movable hydrocarbons Independent of lithology and porosity Inherent compensation for shaliness Best iJSed with R,,, or resistivity overlay Restricted to fresh mud (R,,,> R,)

5 Cyberlook or Prolog Requires deep Resisitivity, Density, Neutron, and GR (or SP) Provides shale-corrected water saturation and porosity Indicates gas if present

* Shows moved hydrocarbon if an R,, log is available Excellent for picking Z O I - I ~ S to test or sidewall core


‘A. Poiipon, C. Clavier, J. Durnanoir, R. Gayrnard, and A. Misk, “Log Analysis of Sand-Shale Sequences - A Systcmatic Approach,” J. Pet Tech. (July 1980). pp. 867-881.

“J.R. Ratliff, W.H. Throop, F.G. Williams, and J.B. Iiall. “Applications of thm! Sarabaad ,%nd- <pes in PWS Syniposiurn Transactions (May 1971).

“A. Poupon, W . R . Hoyle, A.W. Schmidt, “Log Analysis in Formations with Complex Lithologies.” J. Pet. Tech. (August 1971), pp. 995-1005.

‘‘A.W. Schmidt. A.G. Land, J.D. Yunkcr, and E.C. Kilgore, “Applications of the Coriband Technique to Complex Lithologies.” SPWLA Loggirig S y ~ n p < i s i t m Transaction$ (May 1971).

I3G.R. Coatec, R.P. Schuize. and W.H. Throop. ”VOLAN - .4n Advanced Computational Log Analysis.” SPWLA Logging Symposiuni Transuctin~is (July 1982).

I4C. Mayer and A. Sibbit, “Global, A New Approach to Computer Processed Log Interpretation,” SPE 9341, (Dallas: September 19SO).

d. W.H. igital Shaly Wax- og-De Y Typing,” -June

1982).

Chapter 1 O

RECOMMENDED LOGGING SUITES

Many factors influence the choice, including type of formations. T prior knowledge of the reservoir, hole size and deviation, cost of rig time, and availability of equipment.

To provide a general framework for decisionmakirig, recommended ioggingsuites are listed in Tables 10-1 and 10-2. Talde 10-1 is for conditions of fresh mud (R,,, > 2R,) and medium-to-soft rock (R, < 200 ohms). These are conditions appropriate for Induction logs. Table 10-2 i s for conditions of saltier mud (R,,,, < 2R,,) or hard rock (R, > 200 ohms), for borderline cases where the bit size is large (> 10 in.), or where invasion is deep. These are

In each category three PO welIs, one for development \li

Most wells fall in the development well category. Sufficient logs must be run to distinguish gas from oil, to handle lithology variations, and to cope with shaliness. Basically, this means running the Density-Neutron combination for porosity-iitholog). determination and gas indication.

For exploration wells. particiilarly rank wildcats, all information about siibsurface structure, lithology, porosity, and hydrocarbon saturation is desired. It is important to correlate seismic sections with Sonic-Density synthetic seismograms for selecting offset locations and to obtain as much pore pressure information as possible to optimize the offset drilling. All of this requires a full suite of applicable logs.

The number of logging runs required in each case is shown in Tables 10-1 and 10-2. For each run the order in which the logging tools are listed represents the order in which they are combined in the logging array, from bottom to top. Resistivity tools are always on bottom with porosity tools above if they are run in combination. Logging arrays vary from 30-80 ft in length. This means the first reading for a given curve (indicated as FR on the bottom of the log) may vary anywhere from about 3-70 ft off bottom. If it is important to obtain all curves in a target formation close to bottom, the well must be drilled about 80 ft beyond the target or the logging tools must be run separately.

337

338 ESSENTIALS O F M O D E R N OPEN-HOLE LOG INTERPRETATION R E C O M M E N D E D LOGGING SUITES 339

TABLE 10-1 R E C O M M E N D E D L O G G I N G SUITES F O R M E D I U M - T O - SOFT R O C K , FRESH MUD

1 lnfili Weils One run : IndiJctioriiSFL-Sonic

6 curves' SP, lLc,, SFL, t, R,,,, Tension F a s t (5,000 ítihr!; no pads to stick ,4dequate iri clean formatioris when lithology known inadequate in unconipacted shaly, or variable lithology formatioris Unreliable gas iridication insufficient for cornputer processing

2. Development Welis @ne run : Ouol InductioniSFL~Density-Neutron-GR

10 curves' SP, GR, IL,, iL,,, SFI., Slow (1,800 ftihr): pads may cause sticking Handles uncompacted. shaly, or variabie iithoiogy formations Excellent gas indication Inadequate in very rough hole Sufficient for computer processing at wellsite or office

dN. CAL, Rbn, Terision

Additional runs Multiple Formation Tester, Dipmeter, or Sample Taker as reauired

3. Exploration Welis Run 1 üual InductioriiSfL-Sonic Run 2 Litho-Density-Neutron-Microlog-Electromagnetic

Pr»pagation-Spectral Gr Run 1: 5.000ftihr

6 curves: SP. ¡La, !L,F, SFL. t. R,, Run 2 . 1.800 ftihr

13 curves CAL, GR, U, Th. K. P,, @:, pN mi;, MINV, MNOR. Terision Handles variable lithology. shaliness. or rougti hole Excellent gas indication Movable oil determiiiatiori Depth calibration of seismic Complete data for full range of computer processing

Run 3 : Dipmeter Run 4 : iviultiple Formation Tester Run 5 : Sidewall Sample Taker

TABLE 10-2 R E C O M M E N D E D L O G G I N G SUITES F O R H A R D R O C K O R SALT MUD

I . lnfili Wells Run 1 Dual Laterolog-R,, Run 2 Sonic-Gamma Ray

Run 1: 5,000 ftihr 5 curves: SP. LL,. Ll.,. MSFL. CAL

Run 2 , 1,800 ftihr 2 curves' GR, t

Adequate when lithology well known Inadequate in variable lithology Movable oil determination No gas indication Insufficient for computer processing

2. Development Wells Run 1 Dual Laterolog-R,, Run 2 Litho-üensity-Neutron-Spectral GR

Run 1: 5,000 ftihr

Run 2: 1.800 ftihr

Good lithology determination Movable oil deterniination Excellent gas indication Inadequate in very rough hole Allows log interpretatiop at wellsite or office

5 curves: SP, LL,. LL,, MSFL, CAL

9 curves: CAL, GI?. U, Th, K, P,. +o, <pN, Tension

Additional runs: Multiple Formation .Tester, Dipmeter, or Sample Taker as required

Run 1 Dual Laterolog-R,, Run 2 : Dual Induction SFL-Sonic Run 3 : Litho-Density-Neutron-Spectrol GR

Run 1: 5.000ftlhr

Run 2: 5,000ftlhr

Run 3, 1.800 ftihr

Handles low porosities. variable lithology, rough hole Movable oil and secondary porosity calculable Excellent gas indication Depth calibration of selsmic Allows log computation at wellsite and otfice

3 Exploration Wells

5 curves' SP. LL,. L!$, MSFL. CAL

5 curves: SP, !Lo, IL,, SFL, t

9 curves CAL, GR, U . Th, K, P,. 4rl, oN, Tension

Run 4 : Dipmeter Run 5 : Multiple Formation Tester Run 6 : Sidewall Sample Taker

341 340 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION R E C O M M E N D E D L O G G I N G SUITES

FRESH MUD, MEDIUM-TO-SOFT ROCK LOGGING SUITES

These suites are typically run in most of the U.S. and Canada except for the Permian Basin and in parts of South Americ

infili Wells

The Induction-Sonic combination has long been a popular minimum logging siiite and is adequate \vhen formations are fairly clean and are sufficiently compacted. It can be run fast and provides the very useful R,,, quick-look hydrocarbon indication. Fig 10-1 is an example, with SP and coinpiited R,, in Track 1, the resistivity curves in Track 2. and the travel time and tension curves in Track 3. The R,, curve clearly shows one hydrocarbon-bearing zone about 6 ft thick at level F. The R value from zones

excursions to long travel times.

Development Wells

For de~eclopriient wells the Sonic tool is replaced by the Densit! -Neutron combination. All logs can itil! h obtained i n one rim. but logging spced is about three times slower than 1% ith Sonic. The ad\yantages of Density- Neutron are gas indication. independeiice of porosity determination to lithologj. variation?, and abilit] to apply shale corrections.

Fig. 10-2 is an example of the indicated logging suite The curves SP, GR, H,,, arid Ap are in Traek 1: the resistivity curves are in Track 2: and the porosity curves are in Track 3 .

The R,,, curve shows a single hydrocarbon-bearing interval 10 ft thick at zone C. The R, is about 0.05 ohm-m (zoneD) so water saturation at level C averages do OXO 4 or0.35. The hydrocarbon is oil, a$ indicated by the lack of crossover of the Density-Neutron curves. The high porosity of zone C, 24 9 , coupled with the fact the sand appears to be very clean-since GR and SP are low and D and N are overlaying-iiidicates a high permeability for that zone. The high R,,, values at zones A, B, and E should be disregarded because they again are caused by lignite streaks.

___

Exploration Wells For exploration wells two basic logging runs are required. The Sonic log

is needed in addition to the Density-Neutron for seismic depth calibration n c T;b eh

f I I

Yg. 1 O{ 1 Simultaneous InductionSonic log (courtesy Schlumberger)

7

342 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION 343 R E C O M M E N D E D L O G G I N G SUITES

-

A

B

8 4 0 0

8500

Fig. 10-2 Simultaneous Dual Induction-S-FL-Density-Neutron-GR log suite in Gulf Coast sand-shale series [courtesy Schlumberger)

shale alteration are expected, the I.ong-spacd Sonic sliould be run in place of the BHC. The Litho-Density log is desirable for lithology identification, particularly in gas-bearing formations. The spectral CR aids in identifying clays and in distinguishing abnormally radioactive sands or dolomites frorir shaly formations. The EPT-Microlog combination allows estimation of movable oil and pinpointing of permeable zones. If mud cakes greater than Yi in. in thickness areexpected, the Proximity-microlog combination should be substituted for the EPT-Microlog. This would require a third basic logging run, Lvhich may be desirable even with the EPT-Microlog because of the large number of curves to be recorded when it is run with Litho- Density, Neutron and Spectral GR.

Auxiliary runs \vith the Dipmeter, Multiple Formation Tester, and Side- \val1 Sampler are indispensable i n important exploration \veils. Dipmeter logs provide structural and stratigraphic information. Pressure data from the RFT can define reservoir continuity as soon as other \veils in the reservoir are drilled. Sidaval1 cores provide lithology and permeability information.

H A R D - R O C K , S A L T - M U D LOGGING SUITES Hard-rock, salt-mud logging suites are typically run in tlie Perniian

Basin of the U.S. and in areas of the North Sea, North Africa, and the bliddle East \vhere 1 o ~ ' porosities or l«\v \vater saturations result in R, values well above 300 ohni-ni. Generally, these conditioiis occur in carbonate reservoirs. In the Eastern Hemisphere boreholes tend to be large arid H,,,,/R,, ratios lo~v , u.liicli also favors use of J~,aterologs.3

Infill Wells For infill ntl ls the Sonic log along with the Laterolog-Ii,,, combination is

sufficient i f forniatioiis are clean and lithology is \veil known. Two runs are rcqiiired since the two tools are not yet combinable. I iowver , if forniatioiis are not clean or litholog!. is quite \.arialile from well to Lvell, the de\&p- nieiit \vel1 suite niay be preferal>le. A case has been d o c i~niented \ \ , h e sprctral GI1 iriforriiation obtained on new irifill wells led to a 55%) increase i n net pa!' of the restvoir on reanalyzing the older logs.'

Inf i l l \veils i i i i i ) . ofteri be drilled preparator). to waterflooding, in this case kiio\vledge of inter-well communication is important so pressure measurements with the híiiltiple Forniation Tester are desirable.

Development Wells TWCI basic logging runs are required, the first with the Dual 1,aterolog-

RXo tool ant1 thesecond with porosity tools. For the latter the Litho-Density-

344 ESSENTIALS OF M O D E R N OPEN-HOLE L O G INTERPRETATION R E C O M M E N D E D L O G G I N G SUITES 345

Fig. 10-3 Dual Laterolog-R,,-Density-Neutron-GR log suite in Middle East carbonate series (courtesy Schlumberger)

Neutron-Spectral GR combination is recommended in place of the conventional Density-Neutron-GR. The Litho-Density log is needed for lithology identification, particularly in gas-bearing carbonates. The spectral GR aids in distinguishing radioactive dolomites from shales.

Fig. 10-3 is an example log suite run in a Middle East carbonate series.

ally low readings such as those at 205 and 250 ft (relative depths) correlate with density readings of 2.9-2.95 glcc, indicating those intervals to be anhydrites.

The Rt values are 1,000-2,000 ohm-m, in which range the Induction log would be extremely inaccurate. The ratios RLiad/RLids and R,,,IR,, average about 8 and 100, respectively, indicating R, = 1.3 R,,, and that invasion diameter is approximately 30 in. (Fig. 4-18). The R, for the formation is known to be 0.013 ohm-m and to be 0.045, giving R,,íR, = 3.5. Quick- look water saturation between levels 7 and 8, where R,, = 10 and R, = 1.3 x 1,000 = 1,300 ohm-m, is given by Eq. 4 .7 as

Neutron-Density crossover indicating the hydrocarbon is gas. This being the case. porosity at level 3, for example, is given by Eq. 5.13 as

The R, for level 3 is 1,000 x 1.3 = 1,300 ohm-m. Applying Archie's equation 2.8 with c = 1.0 gives

-

This verifies the quick-look value. Both values are abnormally low, probably because cementation and saturation exponents of about 2.5 are more appropriate than 2.0 for these formations; they would give S,, = 0.045, a more realistic figure.

Between levels 7 and 8 the Density and Neutron curves overlay. This interval could be interpreted as liquid-filled limestone. However, since

346 ESSENTIALS OF MODERN OPEN-HOLE L O G INTERPRETATION

ttiere i h gits ;it)o\.ts aiid I)t.lon~ t l i c ~ i o i i t ~ : i i i ( 1 no inter\ c~iiiiigslialc 1);irric.i.s t,sist. tlie iiitc-r\.al is I)rol)ahl!. gns-be;ii.iri< d(i1oiiiitti ‘ I ’ l i c b P, ciirvv of tlic 1,itIio- L>eiisit). log \voulJ rew)l\,e t l ie anihigiiit! iis i n Fig. 5-23 for a hiriiilar case.

Exploration Wells

In exploration \\.ells \\.liere h t l i li’,tlroL.arl)»ii-~>eariiip zones of ver!. high resihtivit!. anti u att,r-imiririg zoncs of cliiite Ion, resisti\ it!, are encountered. it is desirable to run both 1,aterolog and Induction tools. The former n i l 1 read more corrrctlj. in the li!,drocarboii-l)eariTig zones; the latter. more correctll, in tlie \ ~ , ~ i t e r . . l ~ r a r i I i ~ ~ o I i ~ ~ (\\liere R, niust be determined). This is illustrated in Figs. 10-4 and 10-5. In tlie hydrocarbon zone A of Fig. 10-4, the IL,, reads a factor of two lo\ver than tlie IJLc1. In the n,ater-braring zune B of t’ig. 10-5. the I L , reads too high I ) ? a similar factor. An extra logging run fur the DIL-SFL is not rcyiiired since it can be run in combination with the Sonic:. Once K, value5 are estahiished, thcre is no longer the need to run the Indiictiori. I n any event a total of tlirer runs is required for the basic logs.

In the case of tlie porosity tools, the fortliclornirig I h i l Porosit!. Coiiipen- sated Keiitron, described in Chapter ,5, sfiould be run along wit11 the Litlio- Density and Spectral G H , as soon ab is a\xilable. It will iind«ul)tedi!~ provide hetter porosit!. values in tight doioniitic carbonates. The Sonic log i s iiiiportant for secondary porosity iníorinatioii as well ;IS seismic tie-in.

Additional logging runs \vith the Dipmt,ter, h~lultiple Formation Tester, and Sample Taker are required for the sanle reasons as for soft rock exploration wells. \$’liere detection of fractures is important, t h e Fracture Identifi- cation Log obtairitd with the Dipineter tool sliould also I)c run.’

SPECIAL SITUATIONS

I ~ ~ c a l lwrthliole or forrnation condition> often clictatcs \wiations i n iisiial logging suites. ‘í‘\\v c a m are wortti!. of special riiention: that of oil-base miid :ind of heay!.- mineral-iiearing íorn,ations.

Oil-Base Mud

Oil-hase or iri\,ertr.d-oil-t,riiiilsioii niiids are uscd i n some areas to improve drilling efficiency and to iriaintain good liolc conditions, particularly where swdling shales and high temperatures post: drilling prob:enis. Such niuds are electrically nonconductive and, therefore, preclude running any type of resistivity curve except the Induction log. Therefore, even with high formatiori resistivities the Induction log i s the only choice. Furtunately, borehole arid invasion corrections are at a niinimuin with nonconductive miid.

R E C O M M E N D E D L O G G I N G SUITES 347

Density. Seutron, and Sonic logs all perforrii quite well in oil-base mud so the choice bet\veeii thein can be niade on the basis of the guidelines previoiisl!. outlined. I-iowevei, the Electromagnetic Propagation Lo% is not apl)licablc for reasons esplained ir1 chapter 5 .

_____________ _______-___

RESISTIVITY -~

w m s m 2 / m LATtROLOG DEEP

LATEROLOG SHALLOW

MICHOLATEROLOG

INDUCTION 5FF40 --------- 1 10 100 1000- - r

,ILd

-c LLd

Fig. 10-4 Laterolog showing correct resistivity in hydrocarbon-bearing zone (courtesy Scnlumberger)

I 348 ESSENTIALS OF MODERN OPEN-HOLE LOG INTERPRETATION ’

Interpretation-wise, the standard Archie equation 2.8 can be applied to the uninvaded formation. However, a significantly different approach to estimating the movable hydrocarbon is required because mud filtrate invasion is that of oil rather than water.6

SPONTANEOUS POTENTIAL

MILLIVOLTS

10 - Y+ -I-

RESISTIVITY

OHMS M ~ / U

LATEROLOG DEEP --------- MlCROLATtROLOG

INDUCTION t?FFdO----------

? 1 10 100 1000 ,

?

RECOMMENDED LOGGING SUITES 349

Heavy M i n e r a l s Heavy minerals, particularly pyrite (FeS,) and siderite (Fe,CO,), have

been found in important oil-producing reservoirs in Alaska, the Xorth Sea,

the frequency of the log measurement, the more the log is affected. Conse- quently, Laterologs, which are recorded at frequencies below 1 kHz, are preferred to Induction logs, which are recorded at 20 kHz.

In addition. Density logs read abnormally high densities in pyritic forrnations, which leads to porosities that are too low. In some areas such as Prudhoe Day, this forces fallback to the So~iic log as the most reliable porosity indicator.’

‘D.P. Edwards, P.J. Lacour-Cayet, and J. Siiaii, “Log EIaluation i n Wells

‘C. Clat ic-, A . Eieim. and C Scala, “Effect of Pyrite on Resistivity and Other Drilled with inverted Oil Ernitlsion hliid,” SPE 10206, San Antonio (October 1981)

Logging tvleasrirement~.” SI’LCZA Logging Sympos ium Srunsactionr (June 1976).

Fig. 10-5 Induction log showing correct resistivity in a water-bearing zone (courtesy Schlumberger)

m

L? 3

o

co c:

352 INDEX INDEX 353

Compensated neutron logging, 115-135, 192; depth of investigation, 122-123; vertical resolution, 122-123; data presentation, 123-125; environmental correction, 125-127; gas effect, 126, 128-129, 131-132; combined density-neutron interpretation, 128-131; liquid-filled

fled 132-135

Compensated sonic log tool, 96 Compensated sonic logging, 139-159, 199-193: borehole compensated

log, 141-146; vertical resoliition, 142-143; penetration depth, 142-143: data presentation, 143-144: noise spike, 143, 145: cycle skipping. 133, 145-147: interpretation, 146, 148-158: liquid-bearing formation, 148-151; lack of compaction, correction for, 151-153: travel tinie- porosity Conversion, 153-154; gas-bearing formation, 154-156; secondary porosity, 156-157; hole enlargement effect, 157-158; formation alteration effect, 157

Contact log tool, 68

Curve shape, 40-43 Cutting, 1 Cyberlook log analysis, 317, 327 -331. 335

D Deep invasion (logging suite). 337 Density log tool. $15 Density logging, 97-1 15 Density-neutron logging, 95-96, 128-131 Density-resistivity crossplot, 202 Development well (logging suite), 337, 340, 343-346 Direct phase deterniination. 165 Dirtyishaly formation. 7, 35-36: log interpretation for, 40, 53, IS!).

227-266 Dispersed clay method, 236, 262 Dispersed shaly Tand, 233 Dolomite, 5-7, 120,211

Drawdown permeability, 290-292 Drilling disturbance (of formation), 8-14. See also Formation damage Drilling process, 8

Dual induction logging, 70-76. See also Induction logging Dual laterolog. See Laterolog Dual porosity compensated neutron logging, 136-138: application, 136 Dual porosity neutron log tool, 96-97, 136 Dual water model, 236? 242-243, 247-252, 257-261

E

Effective porosity, 227, 252-253 Electric potential, 35-37. See also Spontaneous potential logging

ical conductance. See Electrical conductivity, 17, 70,229, 243, 245 Electrical property. See Electrical conductivity and Electrical resistivity Electrical resistance, See Electrical resistivity Electrical resistivity, 2, 17-24, 26-29, 36, 43-50, 63, 227 Electrical survey tool, 65-70 Electrochemical potential, 38 Electrode array, 65. See also Electrical survey tool Electrofiltration potential, 39 Electrokinetic potential, 39 Electromagnetic propagation log tool, 68, 96-97 Electromagnetic propagation logging. 170-173, 177-191, 194:

application, 171-172; sensor array, 172-173; vertical resolution, 177-178; data penetration depth, 177-178; borehole effect. 177-178; data presentation. 179-150; interpretation, 180-183; travel time conversion to porosity, 180-183: flushed-zone water saturation, 183-184; moved oil estimation, 184-187; heavy oil detection, 187-189; hydrocarbon indication, 189-190; shale effect, 189. 191

Electron density, 102-104 Empirical relations (permeability), 279-284 Environmental correction, 125-127 Equivalent porosity, 120 Excess conductivity, 229-230, 237 Exploration well (logging suite), 337, 341-343, 347

354 INDEX INDEX 355

F

1 ~ 1 0 ~ ~ relations, 267-269. Sce ulso Empirical relatiori\ (perrnea1,ility) Fluid property, 1 Fluid sampling arid analysis, 301-306, 313-311 Flushed-zone log tool, 68, 88 E'ociised logging, 76-82: borehole effect., 77; bed thickness effect. 77. Scc

Fooiised survey tool, 65-68 Formation area, 21 Forination damage, 8-14, 26-29 Formation evaluation, 22-26, 230 Forniation factor, 19 Formation pressure, 267, 307-31-1 Formation thickness, 21, 75-76, 20!)--300 FOR,, tool, 68 Free water, 232 Fresh mud (logging suite), 337-338, 340--313 Fresh mud logging tool, 66-68

also h1icrosI)hericallq. focused logging

G Ganiiiia ray logging, 35-36, 50--61: response to different forrnations,

51-52; depth of penetration, 51: vertical resolution, 51; borehole effect, 51; 53; shale deterniination, 53; spectral gari-iiiia ray log, 53-57; statistical fluctuation, 57-60. Sec. al,vo Spontaneous poteiitial loggirig

Camnia spectroscopy, 223 Gasoil ratio, 30&306 Cas-oil-water contact, 30Cl-310 Gas saturation. Sc>c' li!drocarl)on satriratiori Cause, 308-309 Ciiard log tool, 68 Cypsiim, 21 1

H

Hard rock (logging suite), 337, 339, 343-347 Heavy mineral (logging suite), 349 Iieavy oil detection, 187-189

€iingle plot. 200-207: sonic-resisti\,ity crossplot. 202-00-4; deiicit!.- rtsistivit!, crossplot, 201: nio\.able oil crossplot, 204-206

€I!.drocarlion (A\.aliiation. 1, 11-03, 89--93, 18<1-.190, 317-336 Ii!.drocart)on saturation, 6 ; 13, 17-18, 20, 25-46. 30-31 I I "d rocar lion t r acc . 1 lI!.tlrocarl)oiis in plncci (calculatiun o f ) , 21- 22; 32. 317-336 lI!drogen. 115-11(;, 230

I

Illite, 030--231, 231, 244-245 Iridiictiori log tool. 66-68 Iridiictiori logging, 6$1-76, 92: boreliole effect, 7-1-75: hed thickness

Iiifill \vel1 (logging stiite), 337, 310. 343 Iiictr~iriieiitation, 2 4 In\.adrd zone, 8-10. 12-14, 63 liivasion ( o f foriiiation), 8-14, 26--29, 32-33, 63, 68: depth of, 10-13, 337

effect, 75-76; sondc error, 76

K

K'ioliiiitc,. 230. 237. 211-2-i5

1

1.xninattid shal!~ sand, 231-233 I .attr;iI cxlectrodc array, O5 I. ;i terolog log gin g. 69 -70, G-85 . 9 1 --!O : dcpt h of in\c:si igat ion, 06-87 :

i.aterolog tool, Mi, (is I .inirtorie. 4. 6-7, 120, 21 1 1 .itho-dciisit)~ log tool, O(i-07 Litho-density logging, 108-115, 192: clepentience on lithology, 110-1 11;

vertical resolution. 87: boreholt! effect, 57-88

penetration depth, 11 1 ; borehole effect, 110; example, 112-114; interpretation, 114-1 15

Litho density-neutron method, 11 1, 217--218, 220-223 Iithology, 1, 110-111, 114-115, 166, 169

357 INDEX INDEX 356

Logging suite, 337-349: fresh mud (induction), 337-338, 340-343; medium-to-soft rock (induction), 337-338,340-343; salty mud (Laterolog), 337,339, 343; hard rock (Laterolog), 337, 339,343; hard rock (Laterolog), 337, 339,343-347; borderline case, 337; deep

343-346; exploration well, 337, 341-343,347; oil-base mud, 346-348; heavy mineral, 349

t we1

Long-spaced sonic log tool, 96-97 Long-spaced sonic logging, 158-170, 193: anomalous triggering effect,

161-164; shear travel-time mcasurenient, 164-167, 170; mechanical property determination, 167-168: lithology identification, 168-169: clay indication. 169; gas effect. 169

300-301; sampling/sample analysis, 301-306,313-314; sample analysis interpretation, 306; gas-oil ratio calculation, 304-306; water cut calculation, 304-306; application of pressure measurement, 307-314

N

Natural gas, 5, 26 Neutron log tool, 95, 117 Normal electrode array, 65

O

M

M-N plot, 211-215 Matrix identification (MID) plot. 21 1, 214, 216-219 Medium-to-soft rock (logging suite), 337-338, 340-343

hlicrolog logging, 170- 177, 194: application, 171; sensor array, 172-173;

Microspherically focused log tool, 68 Microspherically focused logging, 83, 85-86, 88. See also Focused logging MID (matrix identification) plot, 211. 214, 216-219 Mineral identification, 21 1-225: M-N plot, 21 1-215; híID (matrix

example, 174-176; mud-cake thickness estimation, 176-177

identification) plot, 21 1, 214, 216-219; litho-density-neutron method, 211,217-218.220-223; technology trend, 223-225

Minilog tool, 68 Montmorillonite, 230-231, 237. 244-245 Movable oil, 30, 184-187,204-206 Movable oil crossplot, 204-206 Mud cake, 10, 176-177 Mud filtrate, 8, 10, 27 kfultiple formation tester, 267, 284-314: operation of, 285-286; pressure

recording, 286-287; pad sealing, 286-290: permeability from pretest drawdown, 290-292; permeability from pressure huildup, 292-301; formation thickness estimation, 299-300; measuring after sampling,

Obsolescence, 63-65 Office-computed log analysis, 333-334 Oil-base mud (logg Oil saturation. See Overpressuring, 309

P Pad sealing, 286-290 Penetration depth. 51, 71-73, 76, 86-88, 103-104. 111, 122-123,

empirical relations, 279-284; drawdowm. 290-292: pressure buildup, 292-301: radial flow analysis. 294-296; spherical flow analysis. 293-299

Pernieable zone log. 35-62: spontaneous potential logging. 35-50, 60;

Photoelectric absorption, 108, 112 I’ickctt plot, 200, 207-208: porosity-reiistivity crossplot, 207 Polyhalite, 211 Pore space, 17. See a h Porosity Porosity, 5. 14, 19, 21-24, 63. 95, 120-122, 156-157, 211. 227, 230,

Porosity log, 95-97 Porosity overlay, 317,319, 321423,334 Porosity-resistivity crossplot, 207 Potassium, 50. 53, 231

gamma ray logging, 50-61

243-245, 252-253

358 INDEX

Pressure hiildiip, 292-301 Pressure depletion, 290-292, 310-31 1 I’ressure rrieasurement. 292-301, 307-314: pressure hi ldi ip , 292-301;

pressure gauge characteristic, 307-308; high accuracy gauge, 308-309: single well application, 309-31 1; overpressuring. 309; gas-oil-water contact, 309-310; pressure depletion, 290-292, 310-311; niiiltiple \vel1 application, 31 1-314; tool improvement, 313

Pressure recording, 286-289 Probe plugging, 287-289 Producibility prediction, 267-315; multiple formation tester, 267,

284-314; formation pressure, 267, 307-314; permeability, 267, 269-271, 279-284; reservoir fluid sanipleíanalysis, 267, 301-306, 313--314; flow relations, 267-269, 279-284; water saturation, 267, 271-279; sand count, 268

prediction Productive zone, 22-24, 32, 267. S w also Permeability arid I’roducibility

Prolog log anal>.sis, 317, 331-3.33, 335 Propagation effect, 74 Proximity log tool, 68, 88 Pyrite, 11 1, 319

R Radial flow analysis, 294-296 Radioactivity, 50-51, 211, 231 Reservoir fluid sampleianalysis, 267, 301-306, 313-314: interpretation,

Resistance. See Klectrical resistivity Resistivity log, 63-94: electrical survey tool, 65-66; electrical survey log,

65-66; focused survey tool, 65-68; Induction log tool, 66-68; Dual Induction log tool, 67-68; Laterolog tool, 66; fresh-niiid logging tool, 66-68; spherically focused shallow log tool, 68; flushed-zone log tool, 68; hlicrolog tool, 68; rninilog tool, 68; contact log tool, 68; electromagnetic propagation log tool, 68; proximity log tool, 68; salt niud tool, 68; Laterolog-7 tool, 68; Laterolog-3 tool, 68; guard log tool, 68; Microlaterolog tool, 68; FOR,, tool, 68; Dual Laterolog tool, 68; dual guard-FOR,, tool, 68; microspherically focused log tool, 68

304-306

Resistivity or F overlay, 317, 322, 324-325, 334-335 Resistivity-porosity crossplot, 199-200 Resistivity tool, 63-93

INDEX 359

Rock matrix, 4-5, i7.96,211 Rock propvrty, 1-15. See, ulso íi> drocarbori saturatioii, l’ernieability,

Porosit! . u r d \Vater saturation

S

Salinity \.ariable: 18 Salt, 211 Salt-mud tool, 68 Salty miid (logging suite), 337, 339. 343 Sampling. See Fluid sanipling and aiialysis Sand, 4,6-7, 18-19 Sand count, 268 Sandstone, 120, 211 Saturation exponent, 20 Seal failure, 287-2S8 Secondary porosity, 157 Shaleishale formation, 8, 35--36, 49, 53, 189, 211, 227-266. See a h Shaly

Shaliness. Sec Dirtyishaly foriiiatiori niid Shaly sarid Shaly sand, 231-236: laminated, 231-233; dispersed, 233; strrietiiral,

234-236: interpretation model. 236; partitioning, 239-240; porosity, 243-246; conductivity, 245-247: volumetric fraction, 248-252; example probleni, 257-261: interpretatioil metiiod, 261-264

sand

Siderite, 110, 3-19 Sidewall neutron logging, 121-122 Siinandoiix method, 236. 262--2(i5 Skin effect. 71 Sodiiini, 237 Sonde error, 76 Sonic log. 77 Sonic logging tool. 95-96 Sonic-resistivity crossplot, 202-204 SP or R,,,/R, overlay, 317, 324, 326-327, 335 Spectral gamma ray log, 53-57 Spherical flow analysis, 297-299 Spherically foccsed logging, 70, 76-82. See also Focused logging Spherically focused shallow log tool, 68

360 INDEX INDEX 361

Spontaneous potential logging, 35-50,60: behavior over long log, 39-40; shape of curve, 40-43; computation of resistivity, 43-49; shaly sand, 49; anomaly, 49-50; vertical migration of filtrate, 49-50; noise, 50. See also Gamma ray logging

Structural shaly sand, 234-236 Sulfur, 211 Surface area, 238

T

Technology trend (in mineral identification), 223-225 Temperature variable, 18 Thorium, 50,53,231

Well logging data, 4,77-80,88,99-104, 123-125, 143-145, 179-180: interpretation, 32, 80-82,88, 105-107, 114-115,128-131, 146, 148-158, 180-183, 199-226,317-336; clean formation data interpretation, 199-226; shaly formation data interpretation, 227-266

Well logging string, 2 Well logging tool, 2 Wellsite computed log, 317-336: apparent water resistivity, 317, 320,

334; porosity overlay, 317, 319, 321-323, 334; resistivity or F overlay, 317,322, 324-325, 334-335; SP or R,,/R, overlay, 317,324.326-327, 335: Cyberlook log analysis, 317, 327-331, 335; Prolog log analysis, 317,331-333,335

Wyllie relation, 148

U Uranium, 50, 53

V

Vertical resolution, 51, 87, 103-104, 111, 122-123, 143-143, 177-178 Volumetric fraction. 238-252

W

Water cut, 304-306 Water resistivity, 254 Water saturation, 5-6, 14, 18-20, 24-26,30-31, 63, 95, 254-257, 267,

Waxman-Smits model, 236, 241 Waxman-Smitsldual water models relations, 243 Well logging cost, 1

271-279: range of uncertainty, 209-211

Documents

modern_hopen_hole log interpretation.pdf