Log interpretation in horizontal wells

Thesis submitted for the degree of

Doctor of Philosophy
at the University of Leicester

by

Stefan Eric Edward Calvert BSc (Birmingham, 1995) MSc (Durham, 1996)
Department of Geology
University of Leicester

January 2002
 UMI Number: U 164095

All rights reserved

INFORMATION TO ALL U SE R S
The quality of this reproduction is d ep endent upon the quality of the copy submitted.

In the unlikely event that the author did not sen d a com plete manuscript
and there are m issing p a g es, th e se will be noted. Also, if material had to be removed,
a note will indicate the deletion.

Dissertation Publishing

UMI U 164095
Published by ProQ uest LLC 2013. Copyright in the Dissertation held by the Author.
Microform Edition © ProQ uest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United States C ode.

ProQ uest LLC

789 East Eisenhow er Parkway
P.O. Box 1346
Ann Arbor, Ml 48106-1346
To my family
Acknowledgements

Firstly, I thank my supervisors Peter Harvey, Mike Lovell, Roger Samworth and
Jeff Hook. Thanks to Mike, Peter and Roger for first conceiving the idea for the project
and steering me through the niceties of horizontal well logs.
I especially would like to acknowledge Roger’s contribution. He spent many an
hour teaching me the fundamentals of logging tools at the start (and throughout) of the
project and has always been available on the telephone for my stupid questions. Also,
for the job at the end of the PhD, so can put into practice some of my ideas. Likewise,
Peter and Mike are thanked for their continual assistance and support especially through
some more traumatic times whilst writing up.
I would like to acknowledge the London Petrophysical Society and Leicester
University for the funding of my PhD studies. The membership of the LPS are thanked
for their time and support, especially those members whom have supplied data and their
time in discussions relating to horizontal well logs. Special thanks to Jeff for assisting
with the finance and intellectual support. In particular, I wish to acknowledge Ian,
David and others from Oil Company 1, also John, Anchala, Steve and George from Oil
Company 2 for all their help and time. Also, Jim White and Schlumberger for assisting
with permission to reproduce Schlumberger information and images used in the thesis.
The Society o f Professional Well Log Analysts is thanked for their $1500 grant
to enable me to attend the 39th Annual Logging Symposium in Keystone, Colorado
during June 1998. The European Association of Geoscientists and Engineers are
thanked for the £300 travel grant to attend the 61st EAGE Technical Conference and
Exhibition in Helsinki during June 1999.
The staff and students of the Geology department of Leicester University of
1996-2001 are thanked for their friendship especially fellow boreholers.
My family is thanked for their support and help with my spelling and grammar
in the thesis. They deserve credit for being some of the few people who will ever read
the entire thesis and for helping me to become the half-literate person I am, especially
mum who helped me through many reading and writing difficulties over the years.
Finally, thanks to the staff at Reeves for their support and encouragement
through the final stages of writing up whilst working.
Disclaimer

The comments or opinions contained within this thesis are those of the author.
Comments that are referenced directly are those of authors referred to. Any inferences
to equipment performance or views of any specific Oil or Service Companies should
not be drawn from the contents of this thesis.
 Contents

Contents

ACKNOWLEDGEMENTS...................................................................................................................... iii

DISCLAIMER ................................................................................................................................ iv

CONTENTS.................................................................................................................................................v

NOMENCLATURE................................................................................................................................ viii

CHAPTER 1: INTRODUCTION...........................................................................................................1-1

1.1 H orizontal W ell P orosity A n o m a lies .............................................................................................1-1

1.2 A HISTORY OF HORIZONTAL WELLS........................................................................................................... 1-1
1.3 O bjectives and h y po th eses ................................................. ................................................................. 1-3
1.3.1 Objectives............................................................................................................................:................ 1-3
1.3.2 Hypotheses............................................................................................................................................1-4
1.4 T hesis St r u c tu r e ...................................................................................................................................... 1-4

CHAPTER 2: POROSITY: DEFINITIONS, MEASUREMENT AND CALCULATION...............2-1

2.1 P orosity d efin itio n .................................................................................................................................2-2

2.1.1 The definition o f porosity...................................................................................................................2-2
2.2 P orosity measurements : C o r e ............................................................................................................ 2-5
2.3 P orosity measurements : Lo gs ............................................................................................................. 2-8
2.3.1 Density and litho-density logs........................................................................................................... 2-8
2.3.2 Neutron log ........................................................................................................................................2-17
2.3.3 Sonic log............................................................................................................................................. 2-27
2.3.4 Nuclear magnetic resonance log.................................................................................................... 2-36
2.3.5 Resistivity log .................................................................................................................................... 2-39
2.3.6LWD Tools.........................................................................................................................................2-41
2.4 P orosity estimation / calculation ................................................................................................... 2-50
2.4.1 Crossplots.......................................................................................................................................... 2-50
2.4.2 Numerical solutions......................................................................................................................... 2-62
2.4.3 Porosity: Approach usedfor thesis................................................................................................ 2-65
2.5 S u m m a ry ...................................................................................................................................................2-68

CHAPTER 3: HORIZONTAL WELL POROSITY ANOMALIES...................................................3-1

3.1 H orizontal wells and their a pplic a tio n ........................................................................................3-1

3.2 O bserved porosity anomalies in horizontal w e l l s ....................................................................3-3

3.3 P erturbation sources .......................................................................................................................... 3-13

3.3.1 Tool errors ...................................................................................................................................... 3-13

V
 Contents

3.3.2 Borehole conditions..............................................................................................................3-37

3.3.3 Alteredformation conditions.................................................................................................3-48
3.4 Su m m a r y ................................................................................................................................................ 3-74

CHAPTER 4: ANALYSIS OF THE DATA SETS............................................................................... 4-1

4.1 D escription of data s e t s ....................................................................................................................... 4-2

4.1.1 Data set 1 ................................................................................................................................4-2
4.1.2 Data set 2 ................................................................................................................................4-8
4.2 P orosity A n o m a lies ............................................................. 4-19
4.2.1 Porosity anomalies - data Set 1.............................................................................................4-19
4.2.2 Porosity anomalies - data Set 2.............................................................................................4-24
4.3 A nalysis techniques .............................................................................................................................4-27
4.3.1 Porosity algorithms...............................................................................................................4-27
4.3.2 Statistical tests.......................................................................................................................4-31
4.3.3 Spectral analysis...................................................................................................................4-39
4.4 A nalysis of the data s e t s .................................................................................................................. 4-42
4.4.1 Data set 1.............................................................................................................................. 4-42
4.4.2 Data Set 2.............................................................................................................................. 4-82
4.5 M ud Invasion .........................................................................................................................................4-120
4.6 Su m m a r y .................................................................................................................................................4-126
4.6.1 Data Set 1............................................................................................................................4-126
4.6.2 Data Set 2........................................................................................................................... 4-132
4.7 C o n c lu sio n s ..........................................................................................................................................4-134
4.7.1 General data quality criticisms...........................................................................................4-134
4.7.2 Porosity equations.............................................................................................................. 4-137

CHAPTER 5: CONCLUSIONS AND FURTHER WORK................................................................. 5-1

5.1 R ecommended porosity calculation pr a c tic es ............................................................................5-3

5.2 P orosity e q u a tio n s .................................................................................................................................5-4
5.3 Suggested practical rem edies ............................................................................................................5-7
5.4 Suggestions for further w o r k ...........................................................................................................5-8

APPENDIX 1: GEOLOGY OF DATA SET 1 ................................................................................... Al-1

APPENDIX 2: GEOLOGY OF DATA SET 2 ................................................................................... A2-1

APPENDIX 3: EXCEL MACRO LISTINGS.................................................................................... A3-1

BIBLIOGRAPHY....................................................................................................................................B-l
 Nomenclature

Nomenclature

Symbol Units Meaning

Fraction,
* Porosity
Percentage or pu
P cP Viscosity of the fluid

r Hz Ratio of the Larmor frequency to the magnetic intensity

$ MeV Energy lost in a typical collision

$ D(Lime) pu Density porosity in limestone porosity units
$N(Lime) pu Neutron porosity in limestone porosity units

(P e)i Electron density index

Afaex pu Neutron porosity excavation correction
Pa g/cm3 Apparent density
Pb g/cm3 formation bulk density
B Average collision cross-section
<h> pu Density porosity
$Dcorrected pu Corrected density porosity
$Dh pu Density porosity hydrocarbon correction
$DSh pu Shale density porosity
Pe Electron density

Pi cp Formation filtrate viscosity

Pf g/cm3 Fluid density

Ph g/cm3 Hydrocarbon density
Pl g/cm3 Density value measured at the long spaced detector
Pm g/cm3 Matrix density
Mass absorption coefficient
Pm f cp Mud filtrate viscosity
Pmf g/cm3 Mud filtrate density
Fraction,
<f>N Neutron porosity
Percentage or pu

$Ncorrected pu Chart corrected neutron porosity

$ neutronfor the Fraction,

Neutron porosity for the appropriate matrix
appropriate matrix Percentage or pu
<fah pu Neutron porosity hydrocarbon correction
&NSh pu Shale neutron porosity
Poil Surface relaxivity of oil
 Nomenclature

Ps g/cm3 Density value measured at the short spaced detecto

P sh g/cm3 Shale density
P su r Surface relaxivity
A t mSec Travel Time
A tf l mSec Fluid travel time
A tm mSec Matrix travel time
A t sh mSec Shale travel time
Px g/cm3 Density of the appropriate material
a Empirical constant
A Mass number
A, B, C Empirical constants
A1 Aluminium
AmBe Americium and Beryllium
API American Petroleum Institute
B Bams (10-24 cm2)
BHA Bottom hole assembly
BP British Petroleum Amoco ACRO
C Carbon
e electron
G Spatial gradient of the magnetic field intensity
Gd Galdanium
Gfr dynes/cm2 Dry frame shear modulus
GM Geiger-Miiller
GR API Gamma Ray
GRSand API Sand gamma ray count
GR-shale API Shale gamma ray count
Gx dynes/cm2 Shear modulus of the appropriate material
h Feet or metres Formation thickness
He Helium
HPormation per cm3 Concentration of hydrogen atoms in the formation
HI Hydrogen index
hr hour
H w a ter per cm3 Concentration of hydrogen atoms in water
Hz Hertz
/ Gamma ray intensity
ILD Ohm-m Deep induction resistivity
ILM Ohm-m Medium induction resistivity
h cps Initial gamma ray intensity
 Nomenclature

K Potassium
k Constant of proportionality
kppm 1000 ppm One thousand parts per million
KC1 Potassium Chloride
mD Formation permeability
Kfi dynes/cm2 Fluid bulk modulus
dynes/cm2 Dry frame bulk modulus
Ks, dynes/cm2 Grain bulk modulus
Kh mD Horizontal permeability
km kilometres
Kmc mD Mudcake permeability
Kv mD Vertical permeability
Kx dynes/cm2 Bulk modulus of the appropriate material
L cm Diffusion length
lbm/gal Pounds per gallon
Li Lithium
Lm cm Migration length
Is cm Slowing-down length
LWD Logging while drilling
m Metre
m Cementation factor
M(t) nV Spin-echo amplitudes of the radio pulses
M0 nVolts Initial Spin-echo amplitude of the radio pi
MAD Measurement after drilling
MD ftorm Measured depth
Mg Magnesium
Mt nVolts Radio pulse amplitude at time period, i
ml/hr-in2 Millilitres per hour per inch squared
MWD Measurement while drilling
N Avogadro’s number
n Saturation exponent
Na Sodium
NaCl Sodium Chloride
NGS API Natural gamma ray spectrometry
Nh, per cm3 Concentration of atoms (hydrogen)
NMR Nuclear magnetic resonance
NPHI pu Neutron Porosity
O Oxygen
OBM Oil based mud
ODP Ocean Drilling Program
 Nomenclature

P ppm Mud filtrate salinity

P i,P 2 psi Pressures
Pb psi Borehole fluid pressure
PEF B/e Photoelectric Factor
Pf psi Formation fluid pressure
PPg Pounds per gallon
ppm Parts per million
psi Pounds per square inch
pu Porosity unit
Rad Ohm-m Amplitude Resistivity
rb Inches or cm Borehole radius
rf Feet or m Distance to limit of reservoir
RHOB g/cm3 Bulk density
n Inches or cm Radius of invasion
Tmc Inches or cm Distance to mudcake from borehole centre
ROP ft/hr or m/hr Rate of penetration
P ps Ohm-m Phase Resistivity
Rt Ohm-m True resistivity
Rw Ohm-m Resistivity of water
R xo Ohm-m Resistivity of invaded zone
S cm Pore surface area
S Sulphur
SDP Slowing down power
sec Seconds
Sg Fraction or Percentage Gas Saturation
Sk su Hydrocarbon saturation
Si Silicon
SP mV Spontaneous Potential
su Saturation unit
Sw Fraction or Percentage Water Saturation
Rxo Fraction or Percentage Water saturation of invaded zone
t Seconds Time passed
T2 nSec Transverse relaxation time
h (iSec/ft Formation bulk travel times
Tb nSec Bulk relaxation time
Tdg nSec Molecular diffusion time
Te mSec Echo spacing
pSec/ft Fluid travel times
Th Thorium
Ti Titanium

X
 Nomenclature

tm pSec/ft Matrix travel times

Ts nSec Surface relaxation time
TVD ft or m True vertical depth
U Uranium
UTotal Total volumetric cross-section
V,, V2 Inches3 or cm3 Volumes
Vb Inches3 or cm3 Bulk volume
Ve Inches3 or cm3 Empty sample chamber volume
Vg Inches3 or cm3 Sample grain volume
Vmf Inches3 or cm3 Volume of invaded mud filtrate
Vmf Inches3 or cm3 Volume of invaded mud filtrate
Vp Inches3 or cm3 Pore volume
Vpx km/sec Congressional velocity of the appropriate material
Vsh API Volume of clay/shale
Va km/sec Shear velocity of the appropriate material
WBM Water based mud
Z Atomic number
B/e Effective formation atomic number
Ap g/cm3 Density correction
Apbmax g/cm3 Maximum density difference
A^Nex pu or fraction Neutron excavation correction
ApRotate g/cm3 Rotational variation correction of the short spaced detector
vm Kg'1 Mass absorption coefficient
Pmeff g/cm3 Effective mud density
§ Section
A Sensitivity factor
AcouPor psec/ft Acoustic porosity
ADN Azimuthal density neutron tool
AT90 Qm 90 resistivity log
B Constant
B/e Barnes per electron
Ca Calicum
CAL2 cm or in Twin arm caliper log (90° to the density caliper)
CALI DEN cm or in Density caliper
CALLISTO CALibration at Leicester In-Situ Tool Optimisation facility
CDN Combined density neutron tool
CP pu or fraction Core porosity
CPOR pu or fraction Core porosity
cps Counts per second
 Nomenclature

CR cps Count rate

Cs137 Cesium137
D Molecular self-diffusion coefficient
DCAL cm or in Differential caliper
DDS N Drillstring dynamics sensor
DEN g/cm3 Density log or density value
DenPor pu or fraction Density porosity
DRHO g/cm3 Density correction
DT psec/ft Acoustic travel time
DTCO psec/ft Compressional acoustic travel time
DTST psec/ft Stoneley travel time
Dw Dimensionless well spacing
DWS cps Short spaced count rate
EWR Qm Electromagnetic wave resistivity
Fe Iron
Frac. Fraction
GOC Gas oil contact
GPDEN pu or fraction Gaymard & Poupon density values
GPNEUT pu or fraction Gaymard & Poupon neutron values
HC Hydrocarbon
IOC Irreducible oil water contact
LADN/LAD_P pu or fraction LWD average density and neutron porosity
LAP P pu or fraction LWD average density porosity
LARES Qm LWD amplitude resistivity
LARHOB g/cm3 LWD average density porosity
LCALI cm or in LWD caliper
LDRHO g/cm3 LWD density correction
LGR API LWD gamma ray
LMD P pu or fraction LWD maximum density porosity
LMDN pu or fraction LWD maximum density and neutron porosity
LMRHOB pu or fraction LWD maximum density porosity
LNP pu or fraction LWD neutron porosity
LNPHI pu or fraction LWD neutron porosity
LPEF B/e LWD photoelectric factor
LPRES Qm LWD phase resistivity
LRES Qm LWD resistivity
LROP ft/hr or m/hr LWD rate of penetration
MMstb Million stock tank barrels
NPHI LIME pu or fraction Limestone neutron porosity
NPHI S pu or fraction Sandstone neutron porosity
 Nomenclature

OC Oil Company (values)

OC_AcouPor pu or fraction Oil Company acoustic porosity
OCJDenPor pu or fraction Oil Company density porosity
OWC Oil water contact
PCL Pipe conveyed logging
ppb Pounds per barrel
QUAL g/cm3 Density correction
Statistical significance at negative values to 99%
rO.Ol(-)
confidence level
Statistical significance at positive values to 99%
r 0.01 (+)
confidence level
Statistical significance at negative values to 95%
r 0.05 (-)
confidence level
Statistical significance at positive values to 95%
r 0.05 (+)
confidence level
r(L) Autocorrelation function
ROPS ft/hr or m/hr Rate of penetration
Rt Qm True resistivity
RT SHIFT Qm Shifted true resistivity values
SDS cps Standard deviation of short spaced count rate
SGR API Uranium free gamma ray
SQ_HC Square root with hydrocarbon corrections
SQ_P pu or fraction Square root porosity
SQPOR pu or fraction Square root porosity
WAP pu or fraction Wireline acoustic porosity
WCALI cm or in Wireline caliper
WDN pu or fraction Wireline density and neutron porosity
WDP pu or fraction Wireline density porosity
WDT psec/ft Wireline acoustic travel time
WGR API Wireline gamma ray
WILD Qm Wireline deep induction
WILM Qm Wireline medium induction
WNPHI/WNP pu or fraction Wireline neutron porosity
WP HC Wiley & Patchett hydrocarbon corrected
WP P pu or fraction Wiley & Patchett porosity
WPPOR pu or fraction Wiley & Patchett porosity
WPEF B/e Wireline photoelectric factor
WRHOB g/cm3 Wireline density
x cm Source to detector spacing
 Chapter 1: Introduction

Chapter 1: Introduction
1.1 Horizontal Well Porosity Anomalies

Horizontal well porosity log data is often anomalous when compared with
vertical well porosity log data through the same formation, typically resulting in
increased porosity values (notably density), although the first example in this thesis
demonstrates decreased porosity values in the horizontal well examined. Several causes
can be proposed, such as: permeability anisotropy leading to irregular invasion and
variable water saturation above and below the borehole (Woodhouse et al. 1991; Cuddy
et al. 1994), differential stress, micro-fractures and disturbed tool placement (Austin et
al. 1994; Cuddy et al. 1994), debris (Cuddy et al. 1994) and bed boundary dip (White
1991). The well data are the best information available that can shed light on these
porosity anomalies hence there is a need to assess which logs provide acceptable
porosity estimates.

1.2 A history of horizontal wells

The beginnings of the modem oil industry can be dated to Edwin Drake’s
discovery on Sunday 28th August 1859 (Howarth 1997), in Oil Creek, near Titusville,
Pennsylvania, where a 69Vi foot well produced oil. However, the Chinese had been
producing oil for over 2000 years and gas for over 3000 years with wells reaching 3500
feet with only bamboo poles with brass attachments. Drake’s motivation was to produce
lamp fuel, since oil was safer than other available forms of lighting at the time.
However, Drake died in 1880 nearly penniless as the oil price fell from $20 to 10c a
barrel in the two years since he first produced oil, due to over production and the
inability of the oil lamp manufacturers to meet demand.
The first published report of a modem horizontal well was in 1939 (Ranney
1939), but horizontal wells had been drilled in the 1920’s. Again, horizontal wells were
not a new idea and had been used by the Greeks, Persians and Egyptians for water
extraction more than 2500 years ago (Nurmi 1995). In southeast England 7500ft long
horizontal tunnels were dug into the chalk for water extraction (Nurmi 1995).

1-1
 Chapter 1: Introduction

However, it was not until the late 1980’s that horizontal wells were drilled to
any great extent. Since then the numbers have increased dramatically especially
throughout the 1990s [Figure l.l](Nurmi 1995).
Horizontal wells are now commonplace and often form an integral part of
planning reservoir drainage, especially for offshore fields. In the Middle East 80% of all
new wells drilled are horizontal (Nurmi 1995). Increased production from horizontal
wells, as shown in Figure 1.2 (Weber 1999), and the additional benefits for solving
drainage problems mean that horizontal wells are often the method of choice.
Having briefly demonstrated the importance of horizontal wells it is necessary
to justify the motivation for the research presented in this thesis. Horizontal wells offer
not only the potential for increased production, but also provide an unparalleled
window into the geology and extent of a reservoir that is impossible from vertical wells
with limited exposure or seismics with limited resolution, thus enhancing upon previous
knowledge. However, log analysis in horizontal wells is not directly analogous with that
in vertical wells primarily because logging tools and interpretation methods were
designed for vertical wells. The petrophysical community is in a constant process of
developing horizontal well log interpretation methods and this is the principle reason
for the London Petrophysical Society interest and funding of the research presented.

1600
Horizontal wells
1570
1400 1400 ! ! Est.
Outside North America
1200

1000
g Canada
United States

800

600

400

200 145

0
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995

Figure 1.1 Number of horizontal wells drilled by calendar year (Figure 1.5 Nurmi
1995).

1-2
 Chapter 1: Introduction

Oil column

Horizi mtal wells (1.900 m. long)

.Vertical well i

E^Dim ensionless w ell spacing

Figure 1.2 Comparison of typical vertical and horizontal well performance (Figure 3
Weber 1999). Recovery factor is the volume of oil recovered divided by the volume of
oil in place and the Dw, the dimensionless well spacing is well spacing divided by the
height o f the oil column all multiplied by the square root of vertical divided by
horizontal permeability.

1.3 Objectives and hypotheses

1.3.1 Objectives

The aims of the thesis study are:

• To identify horizontal well porosity anomalies.

• To describe the potential causes of the horizontal well porosity anomalies.
• To postulate the most likely cause of porosity anomalies using the data sets studied.
• To propose methods of obtaining the formation porosity from log measurements.

1-3
 Chapter 1: Introduction

1.3.2 Hypotheses

1. Log derived porosity values give poor estimates formation porosity in horizontal
wells, regardless of tool type.
2. LWD density-derived porosity provides the best LWD porosity tool estimate of the
true formation porosity in horizontal wells.
3. Rugose and washed-out wellbore conditions in horizontal wells increases porosity
log values to a greater extent than in vertical wells.
4. Formation fluid affects the magnitude of porosity anomalies in horizontal wells to a
greater extent than in vertical wells.
5. In horizontal gas wells, anisotropic mud invasion increases density derived porosity
values and decreases neutron porosity values, without increasing the apparent
variability in the measurements.
6. Bedding and relative dip affect porosity calculations in wells.

The intention was to investigate all the above hypotheses, but data was not
available for use to assess hypotheses 4, 5 and 6. Despite this useful conclusions were
possible.

1.4 Thesis Structure

This thesis consists of five chapters. Chapters 1 to 3 explain the background and
theory behind the thesis. Chapters 4 detail the analysis of two data sets with conclusions
and a discussion of their limitations. The final chapter provides a summary of the
conclusions of the work presented and suggestions for possible further work.

• Chapter 1: Introduction The introduction provides the background to the work

presented and outlines the structure of the thesis.
• Chapter 2: Porosity: Definitions, measurement and calculations The principles of
porosity calculation from log and core measurements are described and the
problems involved are assessed.

1-4
 Chapter 1: Introduction

• Chapter 3: Horizontal well porosity anomalies A summary of notable literature

regarding horizontal well porosity anomalies highlighting fundamental difficulties
of horizontal well porosity calculation and examples of previous works.
• Chapter 4: Analysis of data sets Details the work to resolve porosity anomalies from
two fields. The southern North Sea data set consists of one vertical and one
horizontal well from a Rotliegend sandstone gas field. The northern North Sea data
set consists of two vertical and three horizontal wells from a Palaeocene
sandstone/shale gas and oil field. Porosity anomalies are identified, several porosity
algorithms are statistically tested against core porosity. Conclusions are drawn from
the results of these analyses.
• Chapter 5: Conclusions and suggested further work

1-5
 Chapter 2: Porosity: Definitions, measurement and calculation

Chapter 2: Porosity: Definitions, measurement and

calculation

Formation porosity can be determined from core measurements or a number of

different logging tools: litho-density, neutron porosity, sonic, nuclear magnetic
resonance and resistivity (Ellis 1987). The tools all derive porosity from different
physical measurements and therefore may provide different formation porosity values.
No technique either core or downhole logging measures porosity directly and therefore
no single method provides the true formation porosity in all situations.
Core porosity is often the most reliable measure of porosity, but has probably
never been measured continuously through an entire reservoir section in all wells of any
field. The only scientific projects routinely attempt to recover continuous core from all
wells has been with projects such as the Ocean Drilling Program (ODP).
Continuous formation porosity is usually derived from the density tool
measurements and is regarded as the most reliable log estimation of formation porosity
(Bedford et al. 1997). In general one or two of the porosity tool logs are calibrated
against core in post-logging processing to establish the correct match with formation
porosity. In formations with 10-15% clay content or more correction for the clay is
required, although the correction factors used are almost all without exception
empirical and petrophysicist dependent (i.e. subjective). The empirical relationship is
usually derived from a comparison between core clay content and well log responses.
Note: clay is called shale by petrophysicists regardless of if the clay is a shale or not.
This chapter describes porosity and its role in formation evaluation. The logging
tool principles and porosity estimation from their logs is detailed. Crossplots and
numerical porosity calculations are explained.

2-1
 Chapter 2: Porosity: Definitions, measurement and calculation

2.1 Porosity definition

Porosity is the void space between solid matter in the rock volume and is
occupied by fluids (Tiab and Donaldson 1996). This capacity of rock to store fluid and
its uses form the basis of the entire oil and gas industry. Accurate and precise
measurement of rock porosity may be better described as an art than a science due to
difficulty of the processes involved. This statement may be justified by the number of
methods used to measure porosity and the expense of obtaining good porosity values.
Accurate and precise porosity values are critical in formation evaluation.
Porosity is often used for many purposes although not always reliably; such as an input
for water saturation, permeability, hydrocarbon in place and economic viability
estimates. There are many pitfalls; the most common and problematic is the presence of
significant volumes of shale within the formation of interest. For example, porosity is
often empirically related to the logarithm of permeability from core measurements
allowing permeability to be predicted from porosity logs. However, depending on the
clay content and distribution within the pores this relationship may be complex and
invalid when applied throughout the entire reservoir. This is often where the experience
of the petrophysicist is critical. A relationship may be invoked based on accumulated
knowledge with no scientific data to back up the specific relationship. The only
justification required in this case is the predictive capability of the relationship is better
than previous methods thus it is frequently difficult to evaluate the validity of such a
relationship on any basis other than the specific reservoir or reservoir interval.

2.1.1 The definition of porosity

Porosity refers to the volume fraction of void space of a rock (called total
porosity), but may be described in numerous ways (Bourbie et al. 1987; Dullien 1992;
Tiab and Donaldson 1996). The importance of the measure of porosity for reservoir
evaluation should not be underestimated. Formation porosity defines the upper limit of
the formation’s capacity to store fluid. Porosity is an essential input for water saturation
and permeability estimates, notable problems being clay content and distribution.
Permeability is often estimated from log-derived porosity using an empirical

2-2
 Chapter 2: Porosity: Definitions, measurement and calculation

relationship from core measurements (Bourbie et al. 1987). However, this relationship
is often invalid, since porosity is only one of many factors that affect permeability such
as cementation, clay content and connectivity. Without precise and accurate porosity
estimates it is impossible to calculate the economic value of any reservoir with any
degree of confidence, provided that permeability is great enough to extract the
hydrocarbons at all.
The structure of the pore systems is of significant interest to petrophysicists,
since many rock properties are controlled by the pore structure. For example, the
effective porosity, or the volume fraction of the interconnected voids, is the volume of
rock through which fluids may flow and thus be extracted. Clay content and distribution
is often a factor in determining effective porosity and permeability. However, the
effective porosity does not account for dead end pores. Although dead end pores are
connected to the connected pore volume, they do not contribute to fluid flow pathways
and thus hydrocarbon production (Bourbie et al. 1987; Dullien 1992; Tiab and
Donaldson 1996).
The pore structure itself is controlled by a number of factors (Bourbie et al.
1987; Dullien 1992; Tiab and Donaldson 1996). Porosity is reduced by an increased
range of the grain size distribution that is dependent on depositional environment,
current distribution and the duration of the sedimentary process (Tiab and Donaldson
1996). Cementation affects porosity by filling pore space with material during
lithification and the circulation of fluids during geological history. Diagenetic alteration
of minerals to clay also contributes to the ultimate reduction of porosity by lowering the
strength of the rock and occupying a portion of the pore space, although clay can
possess micro-porosity. Compaction of rock over geological time closes pores and
forces out fluids from the rock. Increasing the degree of order of grain packing reduces
porosity generally because of increased overburden pressure that may also deform or
crush the grains particularly at pore openings providing accumulation sites for clay.
Geologists classify rock porosity to help evaluate the history of a particular rock
sample (Bourbie et al. 1987; Tiab and Donaldson 1996). Two main categories are used.
Primary porosity is the porosity formed at the time of deposition. Secondary porosity is
porosity formed by diagenetic processes such as dissolution, fracturing and catagenesis
(clay alteration due to elevated temperatures and/or pressures). Primary porosity maybe
intercrystalline - voids between crystals, cleavage planes of crystals and voids in crystal
lattices; intergranular - voids between grains; or bedding planes - voids between

2-3
 Chapter 2: Porosity: Definitions, measurement and calculation

sediments and facies (Tiab and Donaldson 1996). Primary porosity voids can also be
formed by detrital fossil fragments; the packing of oolites; depositional vugs or
caverns; and by living organisms at the time of deposition. Secondary porosity is
formed by dissolution - enlargement of voids by warm fluids; dolomitization - the
replacement of calcium by magnesium from solution; or fractures - openings formed as
a result of structural failure. Secondary porosity can also be formed by saddle reefs,
openings at crests of closely folded narrow anticlines; and pitches and flats, openings
formed by beds during slight slumping (Tiab and Donaldson 1996).
An additional problem is that core porosity measurements typically measure the
effective porosity, whereas log derived porosity generally measures the total porosity.
This difference is usually within the errors of the measurements for sandstones because
the non-effective porosity in sandstones is usually negligible (Tiab and Donaldson
1996). However, evaluating carbonates in which the effective porosity may be 0%, but
simultaneously the total porosity could range from 0-50%, results in gross
overestimates of effective porosity (Bourbie et al. 1987; Tiab and Donaldson 1996). In
this extreme case the core porosity, permeability and repeat formation tester
measurements would resolve the issue because without connected porosity there will be
no permeability and no pressure draw down.
For the purpose of this thesis, core porosity is assumed to provide correct
formation porosity since only sandstones are analysed. Inconsistencies that arise are
generally assumed to result from the logging environment or tool malfunctions. This
thesis is only concerned with the porosity that can be estimated from logging tool and
gas expansion core measurements. In this study, porosity can be estimated using core
measurements and the following logging tool measurements: density, neutron, sonic,
resistivity and nuclear magnetic resonance (NMR). All porosity measurement methods
can only estimate porosity but when several methods agree one may have greater
confidence in the estimate of porosity. Note that resistivity and NMR measurement
derived porosity estimates are not used when analysing data in this thesis but brief
descriptions are provided for completeness.

2-4
 Chapter 2: Porosity: Definitions, measurement and calculation

2.2 Porosity measurements: Core

Core sample porosity can be measured in a number of ways, most of which

involve measuring two of the following volumes; total, pore or solid. The methods
include the direct, gas expansion, mercury injection, imbibition, and optical methods.
Porosity may also be calculated from the solid and bulk density of the sample (Bourbie
et al. 1987). The gas expansion method is the most frequently used method. Mercury
injection and imbibition are also common, although mercury injection leaves the
sample unusable for any other measurements such as permeability. Gas expansion,
mercury injection and imbibition measure the connected porosity. Direct, optical and
density methods measure the total porosity of the sample. Generally, the error in
porosity estimation from logging tool measurements is ±lpu and ±0.2pu for core gas
expansion porosity.

Gas expansion

The gas expansion method uses two chambers connected via a two-way valve
(Tiab and Donaldson 1996). One chamber is filled with Helium gas at ambient
temperature and lOOpsi pressure. A sample is placed in the other chamber and
evacuated. The valve is opened between the two chambers, allowing the gas to expand
into the combined volume. The sample chamber volume, Vh may be calculated using
Boyle’s Law (P1V1=P2V2) knowing the initial, Ph and final, P2, pressures and the
volume of the other chamber, V2. Knowing the volume of the sample chamber when it
contains a sample, V,, and the sample chamber volume when empty, Ve, the sample
grain volume, Vg, may be calculated. By measuring the bulk volume, Vb, the porosity, (f
, may be calculated:

V -V
<!> - - b 8 • Equation 2.1

2-5
 Chapter 2: Porosity: Definitions, measurement and calculation

Mercury injection

Mercury injection is used because most substances are not wetted (surface
affinity) by mercury (Dullien 1992). The sample is immersed in mercury because
mercury will not penetrate the pore to measure the bulk volume. The sample is
evacuated then placed in a mercury filled chamber. The chamber pressure is increased
to force the mercury into the pores. The mercury pressure and volume measurements
allow porosity to be calculated.

Imbibition

Imbibition involves weighing the sample before and after imbibing the sample
with a fluid of known density (Dullien 1992). The sample is immersed in a wetting
fluid under vacuum and, given time, the sample will become fully saturated. The pore
volume is calculated by weight after, minus weight before, imbibing divided by the
fluid density, knowing bulk volume porosity is pore volume divided by bulk volume.

Optical

Optical methods use thin or polished sections of core samples, usually

impregnated with epoxy (Dullien 1992). The sample is assumed to be representative of
the formation. The pore area is epoxy filled on the sections and measured visually with
a microscope or, with the use of a microscope camera the section can be digitised. The
digitised image can then be analysed with appropriate software to calculate porosity.

Density

The density method for a dry sample assumes that the solids contain all the
mass. Knowing the bulk volume and density, the porosity equals one, minus the bulk,
divided by the solid density. The solid density may be calculated from the mineralogy.

2-6
 Chapter 2: Porosity: Definitions, measurement and calculation

Direct

The direct method is infrequently used because the method is destructive. The
bulk volume is measured and then the sample is ground to remove the pores (Bourbie et
al. 1987). The volume of the solids is measured. Porosity is calculated by bulk minus
solid volume, all divided by bulk volume.

2-7
 Chapter 2: Porosity: Definitions, measurement and calculation

2.3 Porosity measurements: Logs

This section describes the tool principles and the typical manner in which
porosity is estimated from logs: litho-density, neutron, sonic, nuclear magnetic
resonance (NMR) and resistivity. The aim is to furnish the reader with an understanding
of the fundamental principles of the logging tools and, in addition, to highlight some of
the difficulties involved in deriving continuous porosity estimates from downhole logs.

2.3.1 Density and litho-density logs

Theory: Density

Litho-density tools measure the formation bulk density and photoelectric factor
of the formation. Porosity is estimated using the following equation:

Equation 2.2

where <j>= porosity, pm - matrix density (estimated or known), pf = formation fluid

density (estimated or known) and p b = bulk density as measured by the litho-density
tool (Bateman 1985). The bulk density is calculated based on the count rates of two (or
three) gamma ray detectors at known spacing from a gamma ray source [Figure 2.1]. A
caliper is used to eccentre the sonde to reduce the influence of mud. The count rates
over a specific range of gamma-ray energies can be directly related to the bulk density
of the material between the source and the detectors (Ellis 1987). High bulk density
means more gamma rays are absorbed and so a low count rate is recorded.
The gamma rays interact with the formation in three ways: Compton scattering,
photoelectric absorption and pair production [Figure 2.2](Tittman 1986). Pair
production is avoided by using a source that produces gamma rays at less than
1.02MeV (positron plus electron rest mass energy), but the photoelectric effect does
affect readings. The influence of the photoelectric effect is reduced to an insignificant
level by windowing the gamma-ray spectrum to cut out most of the photoelectric region

2-8
 Chapter 2: Porosity: Definitions, measurement and calculation

(<100keV). The most important of these effects is the Compton scattering, because it is
related to the atomic electron density and therefore the bulk density.
Compton scattering is the process of inelastic scattering of gamma rays from
orbital electrons [Figure 2.3](Tittman 1986). Electrons gain energy by collision with
approximately 2keV-2MeV gamma rays. The electrons are ejected from their host atoms
and the gamma rays lose some of their energy. After a number of collisions, the gamma
rays are absorbed by the photoelectric effect. The greater the electron density, the more
rapidly the gamma rays are attenuated. Compton scattering is the dominant process
between energies of lOOkeV-lOMeV.

Density Logging
Formation

Hydraulic
sonde “

Skid
Caliper
arm

- G a m m a rays
-.Gam m a ray
emitting source

Figure 2.1 Schematic of a density logging tool (Schlumberger 1999).

XT .CALCITE
\Mo%a4o%
P H O T O - P0R0US
ELECTRIC ' \
0.01
SILICA
0% a 4 0 H
POROUS

0.001
.01 .1 I 10
ENERGY (Mev)

Figure 2.2 Gamma ray mass absorption coefficients over the energy range used by
density logging tools (Figure 1 Tittman et al. 1965).

2-9
 Chapter 2: Porosity: Definitions, measurement and calculation

Compton Scattering

Figure 2.3 Compton scattering (Schlumberger 1999).

In common formation elements, Ca, Si, O, C, Mg, K, Fe, Al, etc., the number of
protons, and therefore the number of electrons, is almost half the total number of
nucleons in the atom (the Z/A ratio) [Table 2.1](Ellis 1987). Since these atoms are a
similar size, and the nucleus contains their mass, the formation density can be related to
the gamma ray absorption response of the formation. The higher the density, the lower
the number of recorded gamma rays. The count rate rates for 2 or 3 detectors spaced at
different distances from the source are recorded and the electron density is calculated as
a function of these count rates.
The formation bulk density, pb is related to the electron density, pe by,

Equation 2.3

Equation 2.4

where (pe) i = electron density index, Z = number of nucleons, A = number of

protons and N = Avogadro’s number (6.02 x 1023). However, one notable exception is
hydrogen with only one proton and one electron (Z/A ratio = 1, not Vi). This does cause
problems so the density recorded is an apparent density, pa. Hydrogen is normally found
in the formation fluids and so an adjustment for the hydrogen effect is applied in the form
of an empirical tool compensation.

2-10
 Chapter 2: Porosity: Definitions, measurement and calculation

Elements and Atomic Atomic True Density Electron Density Density seen by Photoelectric
2Z/A
Compounds Number (Z) Weight (A) (RhOb) Index (Rhoe) the tool (RhoJ Index (PEF)

Hydrogen 1 1.008 1.9841 — ... 0.00025

Carbon 6 12.011 0.9991 ... — ... 0.15898
Oxygen S 16 1 ... — ... 0.44784

Sodium 11 22.99 0.9569 ... — ... 1.4093

Magnesium 12 24.32 0.9868 — ... ... 1.9277

Aluminium 13 26.98 0.9637 2.7 2.602 2.596 2.5715
Silicon 14 28.09 0.9968 — — ... 3.3579
Sulphur 16 32.07 0.9978 2.07 2.066 2.022 5.4304
Chlorine 17 35.46 0.9588 ... ... ... 6.7549
Potassium 19 39.1 0.9719 ... ... ... 10.081
Calcium 20 40.08 0.998 — ... ... 12.126
Quartz 11.78 60.09 0.9985 2.654 2.65 2.648 1.806
Calcite 15.71 100.09 0.9991 2.71 2.708 2.71 5.084
Dolomite 13.74 184.42 0.9977 2.85 2.863 2.85 3.142
Anhydrite 15.69 136.146 0.999 2.96 2.957 2.977 5.055
Sylvite 18.13 74.557 0.9657 1.984 1.916 1.863 8.51
Halite 15.3 58.45 0.9581 2.165 2.074 2.032 4.169
Gypsum 14.07 172.18 1.0222 2.32 2.372 2.351 3.42
Anthracite coal 6.02 ... 1.03 1.4 1.442 1.355 0.161
Bituminous coal 6.21 — 1.06 1.2 1.272 1.173 0.18
Fresh water 7.52 18.016 1.1101 1 1.11 1 0.358
Salt water 9.42 --- 1.0797 1.146 1.237 1.135 0.807
Oil 5.61 — 1.1407 0.85 0.97 0.85 0.119
Methane -- ... 1.247 — — — —
Gas -- — 1.238 ... ... ... ...

Table 2.1 Table of density parameters for common elements and compounds
(Schlumberger 1999).

The compensation formula,

Pfl = !• 07(pe)y- 0.1883, Equation 2.5

is derived from calibration in freshwater saturated limestone (Ellis 1987). In practice,

this formula is also a good approximation in water-saturated sandstones and dolomites.
The apparent density, pa is taken as the bulk density and is used to calculate the
porosity, fusing Equation 2.2.
At a detector, the gamma ray intensity, /, behaves exponentially and linearly
depending on the source intensity, 70, bulk density, pb, mass absorption coefficient, vm,
and the source-detector spacing, x (Samworth 1980):

2-11
 Chapter 2: Porosity: Definitions, measurement and calculation

I = I 0ymp bx •e* VmPhX^. Equation 2.6

This equation has a turning point, but the lowest expected bulk density is that of
water, so the tool is designed so that there is an unambiguous count rate to bulk density
relationship. The detector closest to the source is positioned such that the detector
essentially reads mostly the rugosity and mudcake and not the formation (the job of the
second detector). Combination of the count rates from the two detectors is used to
correct for the presence of mudcake. Typically, the density correction, DRHO is,

DRHO = j (pL - p s ), Equation 2.7

where pL is the density value measured at the long spaced detector, and p s is the density
value measured at the short spaced detector. Therefore, the recorded density value, pb
is,

Pb = i ( V i ~Ps)- Equation 2.8

The value of correction, DRHO, is also recorded to provide a quality control

check on the derived density and porosity values. Corrections may be required for the
mud density and borehole size. The larger the borehole size and the greater the mud
contrast between the mud and bulk density, the greater the reduction in the apparent
bulk density measured, because more mud (low density) is seen by the detector.
Generally shale with bounded water will give an overestimate of porosity as will
natural gas because of its low-density compared with water/oil.

Theory: Photoelectric factor

The photoelectric factor uses the gamma-ray spectrum affected by the

photoelectric effect [Figure 2.4](Ellis 1987). The photoelectric effect is the process of
gamma ray absorption by orbital electrons. Orbital electrons are ejected from atoms by
absorbing all the energy from incident gamma rays. The absorbed energy from the

2-12
 Chapter 2: Porosity: Definitions, measurement and calculation

incident gamma ray provides at least the binding energy of the electron and the
remainder is transferred as kinetic energy to the electron. High atomic masses and
therefore greater inner electron shell binding energies require greater gamma ray
energies to eject electrons from an atom. The photoelectric effect becomes significant
with incident gamma ray energies less than 150keV, but is the dominant process for
energies below about 80keV [Figure 2.2]. The spectrum energy of the detected gamma
rays in the density tool is split into soft (low energy, normally about 40-80keV) and
hard (high energy, normally about 180-540keV)[Figure 2.5]. The photoelectric factor,
PEF, is related to the ratio of the soft/hard part of the energy spectrum and is a function
of the effective electron density in the formation, Zeffi

(z V '6
PEF = . Equation 2.9

The photoelectric factor is therefore an indication of the total chemistry of the

formation, aiding lithology and clay identification (~3.5B/e)(Ellis 1987). Typical PEF
values are given above [Table 2.1]. The high photoelectric factor materials can affect
the count rates in the hard part of the spectrum. The weighting agents barite (PEF=261)
and hematite (P2sF=21.48) in drilling muds are a common source of this problem.
These particular weighting agents have high PEF and density values that in
combination result in apparently high bulk density values being recorded [Figure 2.6].
The high bulk density leads to incorrect (too low) porosity values being calculated.
Where the photoelectric factor is measured, the PEF values are sometimes used in
compensation for the density values and are included in the correction curve. However,
this is not sufficient to compensate fully for the effects of barite and hematite. The
example below [Figure 2.6] illustrates only the response of the short spaced detector.
The long spaced detector response is sufficiently different over the range of PEF and
density values expected in the presence of barite and hematite as to make accurate
compensation difficult to achieve. An additional problem is that often barite or hematite
is surreptitiously added to mud systems by drillers without recording its presence!
Therefore, high PEF values should be seen as indicators of possible problems with the
bulk density values. However, siderite (PEF=\A.l) and other minerals with high iron
concentrations also have a similar effect on the PEF and bulk density values.

2-13
 Chapter 2: Porosity: Definitions, measurement and calculation

Photoelectric Absorption

Figure 2.4 Photoelectric absorption (Schlumberger 1999).

Effect of Bulk Density on the Spectrum

(Short spacing detector example)
Lithology window Density window
Region of photoelectric effect Region of compton scattering
(Pband Z information) (Pb information)

1K

0 100 200 300 400 500 600 700 800 900 1000
Energy (KeV)

Figure 2.5 Energy spectrum of a density tool for bulk density and PEF calculation
(Schlumberger 1999).

2-14
 Chapter 2: Porosity: Definitions, measurement and calculation

Effect of Barite Mud on the Spectrum

(Short spacing detector example)

Lithology window Density window

Region of photoelectric effect Region of compton sca tte rin g
(p b and Zinformation) (Pb information)

6K

5K
Low barite and
4K mud thickness

3K
E ffec t of mud thickness
•and barite concentration

High barite arid

mud thickness

0 100 200 300 400 500 600 700 800 900 1000
Energy (KeV)

Figure 2.6 Energy spectrum of a density tool for bulk density and PEF calculation in the
presence of a barite mud (Schlumberger 1999).

Tool differences

The descriptions given in this section rely heavily on Schlumberger documents,

however the fundamental physics of the gamma ray density and photoelectric factor
measurement has firm grounding in the pioneering nuclear physics of the late nineteenth
and first half of the twentieth century. An excellent and relatively accessible summary of
the background physics and detection systems used in all modem density tools can be
found in (Staub et al. 1953; Harvey 1969).
Several descriptions of tools, processing techniques and discussion of error
sources from a number of different service companies are found in the following
references (Czubek 1960; Wahl et al. 1964; Moake 1991) and only minor changes to
wireline tools have occurred since the 1960s. All service companies measure gamma ray
density and photoelectric factor in the same manner because the physics dictates the
nature of this process. The main differences between service companies occur in the data
collection and processing as detailed in the references given above and not the tool

2-15
 Chapter 2: Porosity: Definitions, measurement and calculation

construction i.e. source type (Cs137), scintillation detectors and the spacings are all
fundamentally the same for both wireline and LWD tools.

Common problems

Some notable exclusions are made from the discussion of the operation of
density tools within the literature. This is the process of the application of the source
and detectors to the borehole. The corrections made to the density measurements in
processing are dominantly for parallel standoff of the pad. The problem of non-parallel
standoff is almost ignored in the literature over the past nearly fifty years and is often
glossed over as an intractable problem on the occasions this topic is broached.
The effects of non-parallel standoff can result in poor density values being
recorded and often little or no indication of a problem is evident from the density
correction curve. To safeguard against this effect Oil Companies plot histograms of
density logs across known intervals within a field or region and note inconsistencies.
However, by suitably altering the construction of pad the density measurement can be
constructed more robustly. As important is the toolstring dressing (bowspring,
cranks...), since imbalance of the mechanical forces acting on the entire toolstring can
impart turning motions leading to non-parallel standoff. Analysis of the individual
detector count rates can also provide an indication of this problem, as can the caliper
log. An under-gauge caliper log may indicate mudcake, but also the tool running across
a chord rather than the diameter of the borehole giving non-parallel standoff.
The photoelectric factor log does not indicate standoff as the mud (water)
provides almost no contribution to the photoelectric factor log unless barite is used as a
mud weighting agent [Table 2.1].

2-16
 Chapter 2: Porosity: Definitions, measurement and calculation

2.3.2 Neutron log

Theory

The neutron tool irradiates the formation with neutrons and measures the
neutron count rates at two detectors at different spacing from the source [Figure
2.7](Desbrandes 1985). A bowspring forces the whole sonde against the borehole wall
and shielding behind the detectors helps to reduce the number of detected neutrons that
pass through the mud column. The ratio of the two detector count rates is proportional
to the hydrogen content of the formation within the volume irradiated by the neutrons,
assuming the hydrogen content of the formation to be within water (or hydrocarbon)
molecules of the formation pore volume (Rider 1996). The count rate ratio,
cps„ea/cpSfar, is approximately proportional to formation porosity, $ where k is the
constant of proportionality (Ellis 1986):

near
Equation 2.10
CPSfar

Porosity derived from the neutron porosity tool is calculated using the following
equation:

neutron for the appropriate matrix Equation 2.11

where Neutronfor the appropriate matrix = neutron porosity as measured by the neutron porosity
tool lithology corrected (Bateman 1985).

The neutron porosity tool measurement is based on fast (»100eV ) neutrons

scattering from hydrogen atoms (Tittman 1986). During the lifetime of a fast neutron
there are a number of processes that affect its passage through the formation prior to
detection; elastic scattering, inelastic scattering, diffusion or capture by a thermal
neutron absorber [Figure 2.8].

2-17
 Chapter 2: Porosity: Definitions, measurement and calculation

A fast neutron (3-4MeV) will scatter (predominantly elastically) randomly

through the formation losing energy until the neutron reaches epithermal energies (0.1-
lOOeV), at which point it may be detected or continue scattering until it reaches thermal
energies (approximately 0.025eV). At thermal energies the neutron diffuses randomly
through the formation and may be detected or undergo capture.

Neutron Logging

Borehole Formation

Far detector

Near detector

Neutrons

Neutron source

Eccentrali2er
(bow spring)

Figure 2.7 Neutron logging tool (Schlumberger 1999).

Life of a Neutron
i o 6H
After many collisions
the neutron's energy
is reduced to about Collisions continue but
id4- 0.025 eV (at room with no appreciable energy
temperature). It is loss. Inthis thermal diffusion
now at the thermal region the neutron interacts
energy level. until it is absorbed or
10' captured by an atom. When
10eV\ an atom captures a neutron
a gamma ray is emitted.
A p p ro x im a te
e p ith e r m a l
io°- e n e r g y re g io n
0.4 ftV \
Capture
A v e r a g e th e rm a l V
e n e r g y 0 .0 2 5 eV

10,-2. i
10 100
T im e ( p a e c )

Figure 2.8 Lifetime of a neutron (Schlumberger 1999).

2-18
 Chapter 2: Porosity: Definitions, measurement and calculation

Elastic scattering between a neutron and a nucleus is analogous to two billiard

balls colliding i.e. dependent on the angle of incidence [Figure 2.9]. From this analogy
it is clear that a collision with a body of equal mass to that of a neutron (like the single
proton in a hydrogen nucleus) will result in the greatest loss of energy. This process is
the most significant energy loss mechanism for neutrons at epithermal energies and
above (Ellis 1986).
Inelastic scattering is a process in which the target nucleus is excited by gaining
some internal energy from an incident neutron [Figure 2.10]. This excitation energy is
immediately re-emitted as gamma rays. The amount of energy remaining is dependent
on two factors, the energy of the incident neutron and the target nucleus. There are
characteristic minimum neutron energies below which a target nucleus will not undergo
inelastic scattering i.e. the first excitation state of the nucleus [Figure 2.11]. The
remaining kinetic energy is shared between the target nucleus and neutron according to
the mass ratio. This process is significant for epithermal energies and above.
Diffusion is a process in which the neutrons move after having achieved a state
of energy equilibrium (Barber 1989). For neutrons this equilibrium state is a
homogeneous isotropic concentration of neutrons of the same energy. On average the
neutrons gain as much energy as they lose in collisions, i.e. no net loss in energy. The
neutrons are approximately in thermal equilibrium with the formation, but not in spatial
equilibrium. The net flow direction is from high neutron concentrations to low neutron
concentrations. In logging, at the detectors, this flow direction is normally from the
formation towards the borehole. This is due to the relatively high capture cross-section
of the borehole fluids compared with the formation. The borehole acts as a sink for
thermal neutrons. Diffusion occurs at thermal energies.
Capture gamma rays are produced from nuclei after absorption of (thermal)
neutrons [Figure 2.12]. The new nucleus is excited by the energy gained mainly by a
reduction in the total binding energy of the nucleus. This energy is immediately emitted
as a gamma ray, or series of gamma rays, the energy being dependent on the structure
of the new nucleus.
Neutrons are emitted from the source, experiencing a number of interactions
during their passage through the formation prior to detection (Ellis 1986). As stated
above, elastic and inelastic scattering are the dominant processes that slow the fast
neutrons from the source to epithermal and thermal energies. This slowing down is
highly dependent on the concentration of atoms (hydrogen) per cm3, NH, the average

2-19
 Chapter 2: Porosity: Definitions, measurement and calculation

collision cross-section, crc and the energy lost in a typical collision, £ This effect is
referred to as the slowing down power, SDP, such that for any particular element (Ellis
1987),

SDP = N h<jc%. Equation 2.12

B illiard B all E ffe c t

(all collisions demonstrated here are head-on)

Little energy loss to N

Little energy loss to N

N P
Maximum energy loss to N

| Kinetic energy of the neutron prior to collision

Kinetic energy passed on to the particle afterthe collision
Kinetic energy of the neutron afterthe collision

Figure 2.9 Elastic scattering, N is a neutron and P is a proton (Schlumberger 1999).

Inelastic Neutron
Scattering

Figure 2.10 Inelastic scattering (Schlumberger 1999).

2-20
 Chapter 2: Porosity: Definitions, measurement and calculation

1000
Inelastic Standards

100
F*

Gamma ray energy (MeV)

Figure 2.11 Gamma ray spectrum emitted by inelastic scattering of neutrons with
particular elements (Schlumberger 1999).

Thermal Neutron Absorption

Excited
nucleus
(Z, A+1)

Prompt
gam m a

U nstable
ground state
nucleus (Z. A +1)
Beta panicle

Exceed
n u cleu s
( 2 + 1 . A+1)

D elayed
gamm a
rays

G round s ta le
n u c le u s
( 2 + 1 , A+1)

Figure 2.12 Thermal neutron capture and gamma ray emission (Schlumberger 1999).

2-21
 Chapter 2: Porosity: Definitions, measurement and calculation

Material (Pure) Computed Sigma @ 20°C Rock Matrix Apparent Sigma**

Quartz (SiOj) 4.26 Sandstone 8 to 10
Calcite (CaC03) 7.07 Limestone 12
Dolomite (CaC03 • MgC03) 4.70 Dolomite 8
Anhydrite (CaC04) 12.5
Gypsum (CaS04 • 2HzO) 18.4 **These are real-world values, not
Magnesite (MgC03) 1.44 computations made from pure materials. These
Iron 193 values vary slightly from the theoretical values
Rock salt (NaCl) 753 because of a range of factors, such as trace

Boron (B) 99,405 contamination.

Cadmium (Cd) 113,315

Fresh water (H20 ) 22
Salt water 30 to 130
Oil 20*
Gas 5
'"Because of the minerals produced with the oil, it is not uncommon for unrefined samples of oil to have measured
sigma values greater than 20 c.u.

Table 2.2 Capture cross-section values for common minerals and lithologies
(Schlumberger 1999).

The slowing down length is typically of the order of 10-20cm (porosity and
lithology dependent), but the detectors are typically 40cm and 60cm away from the
source (Ellis 1986). The slowing-down length/power is not the only factor that is
measured by the detectors. The thermal neutrons normally diffuse within the formation
for some distance before being captured by the detector. Another parameter is required
to account for this. The distance travelled by a neutron between its first reaching
thermal energies and its detection (or eventual capture by a nucleus in the formation) is
called the diffusion distance, typically 10-30cm (porosity and lithology dependant). The
slowing-down length, Ls, and diffusion length, L, are normally combined together to
give an estimation of the average distance travelled by a neutron from the source. This
parameter is the migration length, Lm, and is defined as,

4,=V ^+4- Equation 2.13

 Chapter 2: Porosity: Definitions, measurement and calculation

Assuming that elastic scattering from hydrogen atoms in the formation is the
only variable, the measurement of count rate ratio will be a measure of the hydrogen
index, HI, which is proportional to the porosity, </>.Such that,

Equation 2.14

where HFormation = concentration of hydrogen atoms in the formation and H Water =

concentration of hydrogen atoms in an equal volume of water. Therefore, for a given
borehole diameter the porosity can be calculated [Equation 2.10].
However, the porosity calculation is dependent on lithology, salinity (borehole
and formation), mudcake thickness, borehole diameter, standoff, pressure and
temperature [Figure 2.13]. Other effects include inelastic scattering, the chlorine
absorption, the bound water in shale, the breakdown of porosity calculation at high
porosity (>40pu) and the low hydrogen concentration in zones containing natural gas
(Ellis 1986). These effects are compensated for by the use of tool-dependent correction
charts. However, the rugose and caved boreholes tend to lead to the neutron tool
reading too much mud and overestimates of porosity. The lithology has a significant
effect on neutron porosity measurements [Figure 2.32] because the average collision
cross-section of the composite elements of the formation compounds is significantly
different (Ellis 1986). This is especially problematic where mixed/thinly laminated
formations exist. Hydrated minerals are a particular problem and result in porosity
overestimates (Rider 1996). The most common hydrated minerals are clays, in which
accurate porosity estimates from neutron porosity measurements is impossible. Other
examples include gypsum, polyhalite, chamosite, camallite and kainite. Minerals such
as halite, anhydrite and sylvite have no hydration or porosity and have negative neutron
porosity values! Another problem with shale formations is that they may contain
significant quantities of boron (3-4ppm) which would lead to very high neutron
porosity values (Ellis 1986).
Depth of investigation of the wireline neutron porosity tool at 15pu is
approximately 12inches (30cm) with a vertical resolution of approximately lOinches
(25cm) (Rider 1996). The LWD neutron porosity tools are combined with the density
tool. An LWD tool built into a 6V6inch drill collar has a precision of ±0.5pu in 30pu

2-23
 Chapter 2: Porosity: Definitions, measurement and calculation

formation at a rate of penetration of 50ft/hr (Wraight et al. 1989). Note that the
common size of LWD tools are 43A, 8/4 and 12lA inch drill collars.
General rules are that shale (bound water) will give an overestimate of porosity
but, most importantly, natural gas will give an underestimate because of its low
hydrogen concentration compared with water/oil. However, neutron porosity values in
oils may also be affected by the difference between oxygen and carbon collision cross-
section values leading to slight overestimates of neutron porosity in oils (Sherman et al.
1983).

Tool Differences

The neutron tool descriptions given above are from Schlumberger documents.
The measurement technique is physics based from the nuclear physics of the late
nineteenth and first half of the twentieth century and is consistent between Service
Companies. Accessible summaries of the background physics and detection systems
used in neutron porosity tool design are found in (Morrison and Feld 1953; Harvey
1969). Several descriptions of tools, processing techniques and discussion of error
sources from a number of different service companies are found in the following
references (Tittle 1961; Tittle and Allen 1966; Allen et al. 1967; Arnold and Smith
1981; Scott et al. 1982; Scott et al. 1994).
Several generations of neutron porosity tools have existed, although the most
common modem type is the thermal type as described above. The main differences
between service companies occur in the data collection and processing as detailed in the
references given above. The tool construction does alter between Service Companies to
exploit different aspects of the physics of neutron transport through porous materials
however this does not alter the underlying principles described above.
Neutron porosity logs from different Service Companies can vary significantly
for several reasons. The main reason is that the data processing techniques employed
vary considerably. Companies use polynomial fits to laboratory standards directly,
calculating for porosity and correcting for environmental condition; semi-empirical
methods using theoretical formulae with adjustable parameters that are defined by
laboratory calibration for environmental correction; and near fully theoretical methods
using downhole measurements for environmental corrections based on laboratory
standards. The detailed construction of the neutron tool has bearing on the response.

2-24
 Chapter 2: Porosity: Definitions, measurement and calculation

The effect of these differences is seldom noticeable in reservoir rocks

(sandstone, limestone and dolomite) which are well within the calibration range and
standards used. The log responses in shales can be markedly different and can cause
significant problems when making shale corrections to logs in shaly sand. This often
results from shales being out of the calibration range of the tool. Shale is not generally
modelled or calibrated for since the definition of shale is extremely broad. Shale may
contain a variety of rare earth minerals in minute quantities and variable amounts of
bound water that can have large effects on the neutron response. Shale may alter by
chemical interaction with the drillling mud, thus changing the expected neutron
response.
Additional effects of the toolstring configuration whilst logging are significant.
The neutron log is not as susceptible to standoff as the density log due to the increased
detector spacings used in a neutron tool. Though similar arguments regarding mud
balling/rings (build up of mud on the tools e.g. on the leading edge of the density pad),
turning motions and non-parallel standoff resulting from inappropriate tool dressing are
also appropriate for the neutron tool.
The author is aware of 12pu differences in shales, but matching responses in
sandstones in which two Service Companies were run in the same borehole back to
back. This was explained purely on the basis of differing algorithms used to derive
porosity and environmental corrections.

2-25
 Chapter 2: Porosity: Definitions, measurement and calculation

Neutron log porosity index (apparent limestone porosity)

24
20
Actual borehole size 16
(in.)

Mudcake thickness
(in.)

Borehole salinity
(kppm)

2
I 10
Z 9
Q
Mud weight
(Ibm/gal) 18
16
g 14
12
10
8

Borehole temperature
(°F)

Pressure
(kpsi)

Water-base mud
Oil mud

Limestone
formation salinity
(kppm)

Figure 2.13 Neutron porosity correction chart (Por-14c Schlumberger 1999).

2-26
 Chapter 2: Porosity: Definitions, measurement and calculation

2.3.3 Sonic log

Theory

The sonic tool [Figure 2.15] emits successive pulses of acoustic energy into the
formation from two sources and measures the first arrival times of the acoustic energy
pulses [Figure 2.14] at a number of detectors (Rider 1996). The recorded travel time, At,
is inversely proportional to the formation acoustic velocity within the volume of
investigation. By assuming values for matrix, Atm, and fluid, At/i, p-wave travel times, the
porosity of the formation may be estimated (Wyllie et al. 1956):

(A f-A p
Equation 2.15
K - a/ J '

First M otion D e te c tio n (FMD) and

A m p litu d e M e a su r em e n t

A tim e gate opens at the first

Transmitter firing - T q motion detection point. A peak
search is performed on the
received signal within this gate.
The maximum value of the
Preset detection Level signal within this gate is the
amplitude.

Detection gate starts

5 0 0 mV

First motion is detected when

Transit
Time the recei ver si gnal exceeds the
detection level

Time - m illisecon d s (m sec)

Figure 2.14 First arrival time measurement (Schlumberger 1999).

The basis of acoustic logging is the effect of the formation on the propagation of
elastic waves (Rider 1996). The main component of the elastic energy that is used to
evaluate formation porosity is the compressional wave which generally travels faster and

2-27
 Chapter 2: Porosity: Definitions, measurement and calculation

arrives first, although other components such as the shear, Stoneley, Rayleigh, Love and
guided waves are also present (Tittman 1986). All these modes of propagation provide
information about the elastic properties of the formation.
Compressional waves propagate by the particle vibrations in the direction of
energy propagation [Figure 2.16]. The wave velocity is dependent on the density and
elastic properties of the medium. Compressional waves are always the fastest waves
through a given medium and therefore first to arrive at a receiver. They are also the only
mode of elastic energy to propagate through liquids (Sheriff and Geldart 1995).

Two Transm itters/

Four R eceivers

Figure 2.15 Sonic logging tool (Schlumberger 1999).

Shear waves propagate by particle vibrations perpendicular to the direction of

energy propagation [Figure 2.16](Ellis 1987). They travel more slowly than the
compressional waves and so arrive after them. They are produced at the boundary
between the mud and the formation/mudcake by mode conversion. Mode conversion is
the process by which the energy of an incident wave (the emitted compressional wave
from the transmitter in logging) is converted into a number of different modes (shear,
Stoneley...) dependent on the elastic properties of the formation/mudcake. Note that
only compressional wave travel times are used in this thesis for the calculation of
porosity.

2-28
 Chapter 2: Porosity: Definitions, measurement and calculation

W ave P ro p a g a tio n M o d es

At rest Compressional Shear

Figure 2.16 Compressional and shear wave motion (Schlumberger 1999).

Stoneley waves are surface waves (tube waves) produced as a result of partial
acoustic coupling between the borehole mud and the formation (Tittman 1986). A tube
or borehole surface ripples transversely to the long axis of the borehole. The particle
motion is retrograde (in the opposite direction) and elliptical (one motion parallel and
one motion perpendicular to the interface) to the direction of energy propagation. The
displacement is circumferential, like a ring travelling along the tube. Stoneley waves are
produced by interaction between the compressional wave in the mud and the shear wave
in the solid. As the Stoneley wave propagates along the borehole surface, energy seeps
away. The amplitude of the wave decays exponentially away from the interface in both
the formation and the mud. This energy in the decay is mode converted to compressional
waves and it is the mode converted compressional wave that is detectable through the
mud. Stoneley waves are slower than compressional and shear waves, having a velocity
of 86% to 96% of the shear wave (Sheriff and Geldart 1995).
The transmitters are fired in succession after the pulse from the first has died
away. The acoustic pulses are received by four receivers in the middle of the tool
position so that they are typically 3 feet away from one transmitter and 5 feet away from
the other [Figure 2.15]. This system provides borehole compensated travel times
(Kearey et al. 1991). The acoustic pulse is fired from one of the transmitters and
received by a pair of receivers. The time difference between the first arrivals of the pulse
at the receivers is the interval travel time, At. The other transmitter is then fired and a
second At is measured. If the two A fs are the same then the borehole is assumed

2-29
 Chapter 2: Porosity: Definitions, measurement and calculation

smooth and the sonde axis is parallel to the borehole wall, so no compensation is
required. Compensation is achieved by averaging the two travel times.
There are numerous variations on the basic tool design (Tittman 1986), the most
important is that the array sonic tools [Figure 2.17] provide a full acoustic waveform
recording from receiver array. The main purpose is to obtain the shear and Stoneley
wave velocities that are useful for estimating the elastic properties of the formation and
fracture identification respectively. This sonde has two transmitters 2feet apart at the
bottom of the sonde and two receivers at 3feet and 5feet from the closer of the two
transmitters. This allows borehole compensated ACs to be measured. At 8feet from the
closest transmitter are eight receivers placed at 6inches apart [Figure 2.18]. The
waveform is recorded at all eight of the array receivers for about 2500ps to allow all
parts of the wavetrain to be recorded. These waveforms are stacked vertically to create
a variable density plot [Figure 2.19]. These data are processed in a number of ways,
depending on the application.

Porosity

Porosity calculation is dependent on assuming values for travel time through the
matrix, Atm, and formation fluid, Atfl. The log derived travel time provides the bulk
travel time, At, so porosity, (/, can be evaluated using Equation 2.15. This is Wyllie’s
equation (Wyllie et al. 1956) and is based on empirical evidence from experimentation
(Rider 1996). The formula provides accurate porosity values for clean, compacted
water filled sandstones (very low clay content) and carbonates. However, inaccurate
values are produced in dirty (high clay content) formations; undercompacted and
overpressured units, which require compensation or alternative formulae (Dvorkin and
Nur 1998; Khaksar and Griffiths 1999).
There are numerous alternative formulae for calculating porosity from sonic
logs theoretically or empirically derived. One notable empirical formula is the Raymer-
Hunt-Gardner formula (Raymer et al. 1980):

2-30
 Chapter 2: Porosity: Definitions, measurement and calculation

S o n i c D igital Tool

Sonic
digital
cartridge

Mud At
measurem ent
Sonic
section
1099^9
receiver
Eight ceramic
section 3S ft
receivers 45 ft

8ft Two ceramic

Sonic receivers
logging
aft

4b
sonde

Two ceramic
transmitters
□□0

as/8"

Figure 2.17 Modem sonic logging tool (Schlumberger 1999).

A rra y -S o n ic

Figure 2.18 Array sonic tool (Schlumberger 1999).

2-31
 Chapter 2: Porosity: Definitions, measurement and calculation

A rray W aveform E xam ple

Compression8.1 Shear Stoneley
wave wave wave

J'W'vWhVyVVvvnJi^yyv^iirvwv'vv'-'--^

T im e

Figure 2.19 Array sonic tool waveforms used for mode identification and travel time

calculations (Schlumberger 1999).

(A C - A / )
Equation 2.16

where At = travel time, Atm= matrix travel time and Atji = fluid travel time.

Common Problems

Oil and gas do not normally adversely affect sonic derived porosity estimates
(Schlumberger 1989) because they are displaced by mud filtrate within the volume of
investigation. Acoustic velocities through mud and formation fluids are similar and small
compared with the matrix (Rider 1996). An appropriate fluid travel time is usually
sufficient to correct hydrocarbon effects, except if porosity and gas saturation are high.
Additional problems in rugose and enlarged boreholes are cycle skipping [Figure 2.20]
and slipping (or At stretch) [Figure 2.21](Tittman 1986). Cycle skipping is when a whole
multiple of the signal wavelength is missed. Cycle slipping is when less than a whole
multiple of the wavelength is missed.
Cycle skipping can be accounted for in later processing, because whole cycles are
involved causing large and abrupt increases in travel times. However, cycle slipping is
difficult to account for since the portion of a cycle that has been slipped is unknown. The
usual cause of cycle slipping is poor tool centralisation, unless full waveforms are
recorded these problems can be impossible to correct for, although back projecting the
first motion “sine wave” reduces the effect of cycle slipping.

2-32
 Chapter 2: Porosity: Definitions, measurement and calculation

C y c le S k ip p in g

Detection Level
A
Cycle Skipping ^
V V Transit Time error
when the amplitude of
the first arrival falls
below the detection level

Figure 2.20 Cycle skipping (Schlumberger 1999).

At S tre tch
Eccentered
tool response —
Centralized
s . tool response

Detection Level

A t Stretch Transit Time

reading in error
by this amount

Figure 2.21 Cycle slipping (Schlumberger 1999).

In boreholes where large washouts or extensive invasion alteration is expected

long sonic tools are run (Rider 1996). These tools consist of two transmitters two feet
apart at the bottom and two receivers two feet apart, eight feet from the closest
transmitter [Figure 2.22]. The depth of investigation is typically deeper at 15-20inches,
allowing the long sonic tool to measure the undamaged formation beyond the altered
zone compared with 6-10inches of the standard tools. This can be especially helpful in
gas saturated formations in which the standard tools depth of investigation may be only
2inches compared to at least lOinches of the long sonic tools (Rider 1996). Cycle
slipping and skipping are more likely by the use of the long sonic tools due to the
method of borehole compensation employed [Figure 2.22](Rider 1996).

2-33
 Chapter 2: Porosity: Definitions, measurement and calculation

D ep th D erived B o r e h o le C o m p e n sa te d (DDBHC)
The results from tool positions 1 and 2 are added together to produce
borehole compensated results similar to the BHC method.

Tool Moving Up

TT3'

TTT R1

Zone R2
1
ITT 4
TT3
UT
Tool Position 2
(current data) LT
Transmitter Array

Tool Position 2 records transit times

Tool Position 1
from two transmitters to a single (m em orized)
receiver. The zone of investigation Receiver Array
is the area between the two
transmitters. Tool Position 1 records transit times
from a single transmitter to two receivers.
The zone of investigation is the area
between the two receivers.

(TT3-TT4) + (T T 3 '-T T O
Slowness
4

Figure 2.22 Long sonic tool and borehole compensation (Schlumberger 1999).

Additional uses o f sonic tools

Sonic logs are far more useful as seismic interpretation tools for geophysicists to
tie-in with seismic surveys than as porosity tools (Sheriff and Geldart 1995). This can be
further justified by the increased use of dipole and quadrapole sonic tools.
A dipole tool has transmitters and receivers constructed such that the transmitter
causes the formation to be excited with a “shear-like” motion by compression waves
which “push on” one side of the borehole and “suck in” the other (Chen 1988). A
quadrapole (cross-dipole) tool has additional transmitters and receivers mounted at 90°
to the other set to allow for the calculation of shear wave anisotropy (Patterson and
Schell 1997; Vernik and Liu 1997; Tang 1999).

2-34
 Chapter 2: Porosity: Definitions, measurement and calculation

The dipole (and quadrapole) tools are more efficient in energy transfer into the
formation than compressional tools because the energy transmitted through the
formation via the shear mode is greater than the compressional mode. The
compressional (mode-converted) and shear (and other modes) are recorded and can be
used to calculate seismic attributes related to reservoir properties allowing
geophysicists to create attribute maps of the reservoir (Goodway 2001). One such
seismic attribute can be the residual hydrocarbon saturation which when mapped can be
used to plan infill drilling campaigns (Goodway 2001).
Porosity from a sonic tool is generally regarded as a poor substitute for porosity
calculated from a density tool. The use of the sonic tool, although affected by porosity,
is greater in combination with the density tool as the formation density is a useful input
into seismic attribute calculations. The theory on which the sonic log is based has been
available for many years (Wood 1941) and therefore the methods used to record sonic
logs are common to all Service Companies. The use of the sonic log for porosity
estimation was developed early in its’ use (Pickett 1963). As discussed above many
factors affect the quality of sonic logs, mostly noise sources of one type or another. In a
similar manner to the density and neutron tools, the sonic measurements are affected by
perturbations to the ideal toolstring configuration within the borehole. Non-parallel
standoff is accounted for in the tool design, but noise from the actions of tool jewellery
against the borehole can lead to poor results. Stacking the waveforms from each
receiver provides some immunity to this problem for array and dipole tools, while the
rejection of poor data from receivers improves velocity estimates (Kimball and
Scheibner 1998).
Wyllie et al. (1956) and Raymer et al. (1980) are empirical equations [Equation
2.15 and Equation 2.16] and both are in common use. Other methods including
theoretical methods for calculating sonic porosity, are discussed below in section 2.4.2.

2-35
 Chapter 2: Porosity: Definitions, measurement and calculation

2.3.4 Nuclear magnetic resonance log

The NMR tool irradiates the formation with pulses of high intensity
electromagnetic energy at a specific high frequency in the radio region and subjects the
formation to an intense magnetic field (Allen et al. 1997). The tool measures the
amplitude of the radio pulses emitted by the protons (hydrogen in the pore water or oil)
within the formation as result of the magnetic resonance induced by the tool. The
integral of the decay of the radio pulse amplitudes, M, (relaxation time, 7^), over a pre­
set time period, t, is proportional to the formation porosity (Kenyon 1997):

# = I M t exp Equation 2.17

V 2 JJ

The hydrogen nuclei (protons) within the formation fluids align themselves with
the intense magnetic field (polarisation). The tool pulses polarised radio energy into the
formation, which forces the protons (within the volume of investigation) to precess
about the magnetic field in a plane perpendicular (transverse) to the magnetic field at a
specific frequency, the Larmor frequency.
The Larmor frequency corresponds to the absorption energy between two energy
levels within the proton’s nuclear magnetic spin structure. Upon removal of the radio
energy pulse the protons precess at different rates due to local variations in the magnetic
field (de-phasing)(Allen et al. 1997). To compensate for this effect the tool alternates
the polarisation of the radio pulses (by 180° between pulses). The direction of
precession is alternate and this increases the amplitude of the radio signal emitted by the
protons in the transverse plane, (which is measured by the tool) as they lose energy by
interactions.
The protons lose energy in three principle ways: interactions with the pore
surface relaxation, Ts, bulk relaxation, 7&, and molecular diffusion, Tog (Kenyon 1997).
The transverse relaxation, T^, measured by the logging tool is a sum of these
components in parallel:

2-36
 Chapter 2: Porosity: Definitions, measurement and calculation

1 1 1 1
— = — +— +— Equation 2.18
DG

This may be re-stated as spin-echo amplitudes from the 180° pulse sequence,
M(t), that are actually recorded:

M(t) = M 0 exp(-t/Ts )exp(-t/Tb)exp(-t/TDG), Equation 2.19

where M0 = the initial amplitude that reflects the porosity.

Proton interactions with the pore surface are most important for the calculation
of porosity with NMR logging tools. The other two components are unrelated to the
porosity and are therefore noise as far as porosity estimation is concerned.
The protons transfer their energy to the lattice of the matrix at the pore surfaces
as they relax. The rate of energy transfer is dependent on the pore size, S/Vp (surface to
volume ratio), and the surface relaxivity, psur (ability of the surface to cause the
relaxation of the proton magnetisation):

_L=
r ji P S U T
A Equation 2.20

Thus the larger the pore the larger the signal amplitude recorded because the
protons are able to resonate for longer before relaxation. The surface relaxivity, psur, is a
constant for any particular formation. The ‘constants’ are similar for sandstones, but
different ‘constants’ are required for carbonates and other lithologies. The porosity can
be directly measured by the sum of NMR T2 amplitudes, provided the effect of the other
two components is minimal:

= Z M t exp = 1 M 0 exp Equation 2.21

p JJ

Interestingly, the surface relaxivity, poil, of oil is usually smaller than for water.
Therefore, by inspection of the T2 amplitudes, oil and water maybe differentiated. The
sum of the amplitudes will still provide accurate porosity values.

2-37
 Chapter 2: Porosity: Definitions, measurement and calculation

The bulk relaxation, Tb, is dependent on temperature, T, and viscosity of the

fluid,

Equation 2.22
Tb ~ T

Since the temperature can be measured and the viscosity of water is a constant
at a given temperature, Tb, is usually a constant. The bulk relaxation can become
significant in oils (especially tar), dependent on their viscosity.
In practice, the molecular diffusion relaxation time, TDG, is not usually a
problem for modem NMR logging tools (Kenyon 1997). The molecular diffusion
relaxation, TDG, is due to the inhomogeneity of the magnetic field and magnetic
susceptibility differences between the matrix and fluids:

Equation 2.23

where y = the ratio of the Larmor frequency to the magnetic intensity, G = spatial
gradient of the magnetic field intensity, TE = echo spacing and D = the molecular self-
diffusion coefficient of the fluid. This effect can be important in formations or fluids
containing paramagnetic ions. For example, siderite sandstone formations or oils
containing vanadium.
Modem NMR tools are designed as pad mounted or mandrel openhole wireline
tools (Allen et al. 1997). Both designs are such that they are insensitive to the borehole
and rugosity assuming the effects are within the tolerences of the tool. The pad tools are
focused so that their volume of investigation is sensitive to a depth of l-6inches (2.5-
15cm) into the formation. The tool can tolerate rugosity at least as well, if not better
than, the density tools. The mandrel tools are focused on a cylindrical volume
approximately 6inches (15cm) into the formation and can tolerate rugosity and
washouts close to 6inches (15cm). LWD NMR tools are just becoming commercially
available at present.

2-38
 Chapter 2: Porosity: Definitions, measurement and calculation

2.3.5 Resistivity log

There are a number of resistivity tools that measure the conductivity of the
formation (Schlumberger 1989). The primary control on the resistivity (reciprocal of
conductivity) is the formation porosity. The resistivity of shale free brine saturated
formation, Rt, divided by the resistivity of the brine, RWy is proportional to the reciprocal
of the porosity, f to the power, m (typically 2)(Archie 1942):

Rw
(p = • Equation 2.24

Resistivity is seldom used for estimating porosity since its primary use is for
calculating water saturation, which requires porosity as an input (Worthington 1985).
However, the above equation can be used in water-saturated formations.

R esistivity Cube

Multimeter

1. Apply a known (regulated) DC

voltage(V)across opposite sides
of a 1 cubic meter sample.
2. Measure the current (I) flow.
3. Determine the resistance (R) from
Ohm's law(R = V /I).

Figure 2.23 Resistivity cube (Schlumberger 1999).

The rock matrix can usually be assumed to be insulating. The conductive

medium in the formation is the pore water (Archie 1942). If a cube is 100% water the
porosity is 100% and the formation resistivity is equal to the resistivity of the water
[Figure 2.23]. If rock, an insulator, is added to the same cube the porosity decreases and
the formation resistivity increases. The relationship between formation resistivity and
porosity is to the inverse square. However, the formation resistivity is also dependent

2-39
 Chapter 2: Porosity: Definitions, measurement and calculation

on the salinity, cementation, tortuosity, granular shape and size. This complicates the
relationship between formation resistivity, Rh and porosity, ^(Winsauer et al. 1952):

aRw
R, = —— , Equation 2.25
<f>
where m = cementation factor (usually 2), Rw = water resistivity, a = constant (usually
1). One common variation is the Humble equation in which <2=0.62 and m=2.15
(Winsauer et al. 1952). However, when hydrocarbons are present the proportion of pore
space occupied by the hydrocarbon must be accommodated. The effect on formation
resistivity of the water saturation, Sw is:

n aR»
'= Equation 2.26
<i>

where n = the saturation exponent (usually 2). Therefore, in the presence of

hydrocarbons an independent porosity or water saturation measurement is required. In
practice, the density porosity is usually used for <f>and Equation 2.20 is rearranged for
Sw. The constants a, m, n and Rw are usually derived from core or using a Pickett plot
[Figure 2.24], Note that the main use of the resistivity measurement is to calculate a
water saturation log with an independent porosity log as input.
500 1000 2000 5000 1 0 ,0 0 0

Fr , fo rm a tio n resistivity fa c to r

Figure 2.24 Pickett Plot (Por-1 Schlumberger 1999).

2-40
 Chapter 2: Porosity: Definitions, measurement and calculation

2.3.6 LWD Tools

This section describes LWD density and neutron porosity tools and highlights
the differences between LWD and wireline tools. This section is only intended as guide
since a fair comparison between LWD and wireline tools can only be made with
extensive computer modelling and field examples and is covered to a degree in the
following chapter. The majority of the information in this section was derived from
Schlumberger’s Logging While Drilling booklet (Schlumberger 1995).
Most LWD companies combine the density and neutron tool into one drill collar
[Figure 2.25]. The specifications of LWD density and neutron tools are similar between
Service Companies, Schlumberger’s specifications are given [Table 2.3]. The density
neutron tool is always placed at the top of the toolstring furthest from the bit so that the
nuclear sources can be retrieved by slickline if they become stuck [Figure 2.26].
Consequently the time when measurements are recorded may be several hours after
penetration depending on the rate of penetration and bottom hole assembly.

0.5pu below lOpu 5% of ±lpu at 30pu at 13.2in

Neutron Porosity 0 to lOOpu
reading for 10 to 50 pu 5Oft/hr [15m/hr] [33.5cm]
±0.1g/cm at 2.5g/cm3 6.2in
Bulk Density 1 to 3.1g/cm3 ±0.02g/cm3
at 5Oft/hr [15.7cm]
±0.25units from 1 to ±0.05units at 3units
Photoelectric effect 1 to lOunits 2in [5cm]
5.1units at 5Oft/hr

Table 2.3 LWD porosity tool specifications (Table 6.1 Schlumberger 1995).

2-41
 Chapter 2: Porosity: Definitions, measurement and calculation

Neutron
sensors
Fishing
head Neutron
source

Electronics
Titanium

Density
channel source

Clamp-on Density
stabilizer sensors
W

Power supply
and batteries

Figure 2.25 Schematic of the CDN density neutron tool (Figure 6.2 Schlumberger
1995).

Nominal Nominal
6Vfe- and 7-in. OD 8- and 8 1/4-in. OD
T

CDN tool
20 ft LStabilizer
hr Nominal
I l 8 ft 9- and 9Vfe-in. OD

_ Modulator
_Centralizer
MWD Electronics I _MWD .MWD
34 ft
- D&l sensor
Turbine/
DWOB alternator
sub - Weight and
6ft torque sensors
T

CDR tool
22 ft

I" 7 ft

Motor

£
Figure 2.26 Schematic of typical LWD toolstring (Figure 2.4 Schlumberger 1995).

2-42
 Chapter 2: Porosity: Definitions, measurement and calculation

LWD neutron porosity tools

There are at least three different LWD neutron porosity tool designs in present
use. One design uses 3 He3 detector banks and 1 Geiger-Miiller detector bank with a
centralised 7Curie AmBe source. Another design uses an eccentred source with Li6
detector banks. The final design employs 4 banks of 4 Geiger-Miiller detectors with an
eccentred sources (Hutchinson et al. 1991). All the LWD neutron porosity tool
source/detector spacings are comparable with wireline neutron porosity tools.
The main differences between LWD and wireline neutron porosity tools are
their environmental responses. The environmental corrections required are similar in
magnitude for LWD and wireline apart from the standoff corrections. Due to the
centralised source in the LWD tool, the sensitivity to standoff is much greater than for
wireline. For example, at 40pu a 2in standoff would increase the LWD neutron porosity
value by 12pu, but only 1.5pu to 3.0pu for the wireline values. This is due to the LWD
tool being affected by removal of formation in all directions. The wireline tool is only
affected by the removal of formation in front of the tool. This difference is particularly
important for horizontal wells where LWD tools often experience significant standoff.
The different designs are used to overcome the problem of drill collar iron
absorption of thermal neutrons (Burnett et al. 1990). Thermal and epithermal neutron
measurement (He3 detectors) of apparent porosity show slightly less sensitivity to
formation porosity than capture gamma ray measurement (Geiger-Miiller detectors) in
the LWD logging environment. The iron drill collar radiates the captured neutron
energy as capture gamma rays, the resulting tool measurement characteristics are
dependent on the iron drill collar. However, one advantage is that the collar displaces
borehole fluid that has greater neutron slowing down power, but iron absorbs 9.8 times
more thermal neutrons than water/mud. Thus, LWD neutron porosity tools (all have
some capture gamma ray response) are less sensitive to borehole and formation salinity
effects (Evans et al. 1988). This lack of sensitivity (the iron drill collar reduces the
thermal neutron flux) leads to lower recorded porosity values than wireline tools, where
no washouts are apparent. The effect has been observed to increase in the presence of
high iron concentrations (siderite and glauconite)(Sakurai et al. 1992).

2-43
 Chapter 2: Porosity: Definitions, measurement and calculation

L W D d e n s ity to o ls

LWD density tools vary in design considerably less than the neutron porosity
tools. The detectors are scintillation counters or GM tubes and may be run with or
without stabilisers. The 1.7Ci 137Cs source is always eccentred (Hutchinson et al. 1991).
The detector spacings for the LWD density tools are comparable with those of their
wireline equivalents. Another notable difference is that the source and detectors are not
inline, but offset in the direction of rotation [Figure 2.27].
LWD density tools with scintillation counters are dependent on the effectiveness
of the stabiliser to exclude mud from in front of the detectors. Their response is more
sensitive to enlarged and rugose holes and is different to wireline tools (Allen et al.
1990). The correction required for LWD tools is typically significantly greater than 1.5
times the correction “ available” (far to near density difference) across the range of
densities encountered in logging, thus magnifying any errors particularly affecting the
near detector. The correction required for wireline tools is almost a constant 1.5 times
the correction available [Figure 2.28].
The main reason for the superior quality of wireline density tools is that the
standoff remains small due to the small size of the pad adjusting to the shape of the
borehole. In addition, the photoelectric factor measurement is more challenging in the
LWD environment because the 1 inch steel drill collar of the LWD density tools would
eliminate all gamma rays with energy <100keV; without low Z (atomic number) ports
in front of the source and detectors a photoelectric factor measurement is impossible.
The ports and stabiliser combine to produce a highly forward collimated density
measurement. Along with the rotation of the tool leads to variable standoff during the
density measurement process [Figure 2.29] and requires correction [Figure 2.30].
A statistical approach is used to account for the variable standoff based on the
variability of the recorded count rate:

AV a ria n c e = AV a ria n c e mamirtd ~ & V a ria n c e cxptcled [Equation 2.27]

2-44
 Chapter 2: Porosity: Definitions, measurement and calculation

Figure 2.27 Schlumberger CDN tool with stabiliser (Figure 6.1 Schlumberger 1995).

P true = 271 Q/cm3 limestone formation

Correction required (ptme_ p/s)

Correction available (pte _ p^)

Apparent
Density,
g/cm3

Long-spacing density(pte)

Short-spacing density(p5 5 )

1.5-g/cm3 nonbarite mud

4 Tool
Standoff, in.
16.5 Borehole
Dameter. in.

Figure 2.28 Modelled CDN density response with stabiliser (Figure 6.12 Schlumberger
1995).

2-45
 Chapter 2: Porosity: Definitions, measurement and calculation

where AVariance is attributed to standoff. The standard deviation of the short spaced
detector count rate is used to measure the variance due to its sensitivity to standoff. The
short spaced count rate is sampled every 100msec and after lOsec (100 samples) the
average and standard deviation are calculated (accounted for in the calculations
[Equation 2.28]). The expected distribution is a normal distribution, if the variation is
bigger than expected, then the change in count rate, ACR is such that:

[Equation 2.28]

where SDS - standard deviation of the short spaced count rate and DWS = short spaced
count rate. The maximum and minimum count rates are at one standard deviation,
DWS±ACR, a compromise due to range of borehole shapes possible. The rotational
correction, Aprotate, is based on this difference:

f DWS + ACR'
APmate = ShortSens x In [Equation 2.29]
kd w s - a c r j ’

where ShortSens = the short spaced detector sensitivity. The formation density is
represented by the lowest count rate from around the borehole, DWS-ACR and is used
to calculate the short spaced maximum density. The maximum density observed, Apbmax
is equal to the average plus half the rotational variation correction of the short spaced
detector, Aprotateplus the usual corrections (See Section 2.3.1);

A , = p ,+ & p + B xp ,rotate ’ [Equation 2.30]

where pt = long spaced count rate, Ap= density correction [C(prps), constant, C (>1.5)
times the difference in the long and short spaced densities] and B= constant (0.5 in light
muds and -0.1 in heavy muds).
 Chapter 2: Porosity: Definitions, measurement and calculation

Circular
borehole Collar

Detector -
Window in
collar and
stabilizer
Form ation

Enlarged Standoff
borehole

Gravity

Figure 2.29 LWD density neutron tool in a vertical (top) and horizontal (bottom)
borehole (Figure 6.10 Schlumberger 1995).

E x p e c te d

G a u g e h o le

M easu red
S a m p le s

E n la r g e d
h o le

M ea n

Count R a t e ►

Figure 2.30 Standard deviation rotational correction method (Figure 6.14 Schlumberger
1995).

2-47
 Chapter 2: Porosity: Definitions, measurement and calculation

LWD Caliper

Differential caliper values, DCAL, can be generated by a statistical approach to

the density count rate distribution (Best et al. 1990). The standoff effect, DCAL, being
dependent on the amount of standoff and the formation, pb, to mud density, Pmeff,
contrast;

DCAL —^ X^ rotate
U^ A L ~ ' [Equation 2.31]
Pb~P meff

where A = sensitivity factor is a polynomial [Figure 2.31]. The effective mud density is
used because barite in the mud can make the measured bulk density appear greater than
the actual bulk density.
Inaccuracies in density derived caliper values can arise due to the linear averaging
of count rates in the algorithm (Spross et al. 1995). However, bulk density is dependent
on the logarithm of the count rates. Hole size, hole shape, sample rate and density
contrast affects the density value accuracy up to a 1" stand-off tolerance.

2.5

2.0

1.5
S ta n d o ff

1.0

0.5

0
0 0.2 0.4 0.6 0.8 1.0 1.2

P/Pb ” Pmud

Figure 2.31 LWD density caliper response data (Figure 6.16 Schlumberger 1995).

2-48
 Chapter 2: Porosity: Definitions, measurement and calculation

New LWD tools

New developments in tool design have included an LWD azimuthal density

neutron tool, which is capable of producing a density image of the borehole from 16
sectors around the borehole (Holenka et al. 1995). This new LWD density tool is now
using three ultrasonic transducers 120° apart to measure borehole diameter. A logical
future development would be to produce an ultrasonic image of the borehole from the
measurements to further aid the analysts.
LWD sonic tools are also available and an LWD NMR is in field trials. Soon
there will be LWD measurement type equivalents of all the available wireline tools.
However, it is doubtful if LWD measurements will ever truly match the intrinsic
quality of wireline tools purely because of the adverse conditions under which the
LWD measurements are made. Their primary advantage is still the early, if limited,
data delivery and the potential of measuring the formation in near virgin conditions.

Brami (1991) provides a discussion of the effects of the different LWD tools
offered by several Service Companies.

2-49
 Chapter 2: Porosity: Definitions, measurement and calculation

2.4 Porosity estimation/calculation

2.4.1 Crossplots

This section details common ways in which two or more porosity estimates are
used to calculate porosity by the use of crossplots. The effects of shale and hydrocarbon
on porosity estimates are discussed. It is stressed that particularly in the case of the
choice of shale equations that the choice is preferably based on scientific data from core
measurements. However, often subjective empirical correction factors are used, or a
prescribed correction for equity reasons (to meet regulatory requirements), rather than a
sound scientific relationship between the core and log data. There are a number of
reasons for this approach:

• Firstly at present no logging tool provides accurate measures of clay volume or

distribution therefore many comprises have to be made. Until an accurate clay-
logging tool is developed, clay estimation will remain an art not a science.
• Secondly core data is often very sparse and therefore a core-log relationship may
not necessarily be applicable on field wide basis.
• Thirdly specific relations between the measurable properties of minerals in a core
lab are particularly difficult e.g. X-ray diffraction mineral identification may be in
error by 50% or more (Doveton 1994). These errors are amplified in the logging
environment.
• Fourthly there may be operational reasons for correction factors, such as the use of
different companies’ logging tools and different generations of tools or due to tool
minor malfunctions (e.g. poor calibration).

Frequently linear relationships are used for simplicity in the full knowledge that
the true core-log clay relationship is non-linear. Most of the clay equations used are a
derivative of the generic equations presented below.
It is normal to crossplot porosity logs to improve the porosity estimation
(Schlumberger 1989). The common crossplots are density-neutron (Figure 2.32),
density-acoustic, and neutron-acoustic. Logging company chart books contain

2-50
 Chapter 2: Porosity: Definitions, measurement and calculation

crossplots with the theoretical porosity curves for common lithologies assuming fresh
water or brine is the saturating fluid. Modem logging software includes the porosity
charts from which the appropriate crossplot porosity may be estimated. The crossplot
porosity for the density/neutron crossplot may be estimated by (Bateman 1985):

Equation 2.32

where (j>N(Lime) = neutron porosity in limestone porosity units and (f) D(Lime) = density
porosity in limestone porosity units. Although this method is in common use it is
accepted only as a porosity estimate. Porosity crossplots are effective if the formation
contains a linear mix of two of the three aforementioned lithologies (sandstone,
limestone and dolomite) and are used with reference to theoretical curves for single
lithologies to aid lithology identification. For more than two lithologies more logs are
required as inputs [Figure 2.33].
Clay is the most common group of secondary matrix minerals and effectively
replaces matrix and/or porosity with a material with different properties from that of the
matrix and pore fluids (Bateman 1985). The crossplots can be adapted for shaly
formations (Schlumberger 1989). Using the appropriate matrix values for clay, water
and matrix, a triangle is formed. By equally dividing the triangle into ten percentile
sectors the shale volume and porosity can be estimated (Figure 2.34). In addition, a clay
indicator may be added to this plot, as a third input, by colour coding. For example,
gamma ray values are useful to quantify the relationship between the crossplot clay
volume. Clay is indicated by high gamma ray values, although problems exist with non­
radioactive shales (some clays have very low K and Th contents), radioactive sands
(often K-feldspar) and U rich formation fluids, U-salt deposits in organic matter or
hardgrounds (Hurst 1990). This empirical method is frequently used despite its
drawbacks. The principal drawback is that the choice of shale/clay point is subjective
(Society 2001), therefore implementation is rarely repeatable.
The gamma ray log measures the natural gamma ray radiation of the formation
and can be very useful for estimating the clay content of the formation (Ellis 1987).
Predominant gamma ray emitters in common formations are Potassium, Thorium and
Uranium, Potassium and Thorium being usually associated with clay minerals.
Generally, the total gamma ray count rate is an excellent quantitative clay/shale volume

2-51
 Chapter 2: Porosity: Definitions, measurement and calculation

measurement. The volume of clay/shale, Vsh, may be calculated from the gamma ray
log, GR, by using empirical values for the shale gamma ray count, GRshale) and the sand
gamma ray count, GRsand:

Problems with this clay volume estimate may occur when Uranium ions are
mobile within the formation waters, or when the formation contains significant micas,
feldspars, potassium evaporites, etc (Ellis 1987; Hurst 1990; Rider 1996). Clay
distribution within the formation alters the choice of shale volume equation as this
impacts the volume proportion of the formation calculated that is clay and therefore the
porosity values (Katahara 1995).
Spectral gamma ray tools are able to separate the contributions to the total count
rate of the three components. The use of the separate components can aid lithology
identification. Thorium is usually associated with heavy minerals within clay and is
usually the most reliable clay indicator especially when used in conjunction with
Potassium (Hurst 1990). Potassium is associated with feldspars (arkose sandstones,
granite) as well as clay minerals. Uranium is mobile, but may also associate itself with
organic matter, often found within clay rich sediments (Hurst 1990). The spectral
gamma ray can be invaluable in certain circumstances when it is impossible to
differentiate lithology by any other means.
Where three or more lithologies are present additional information from other
logs is required for accurate porosity estimation (Doveton 1994). In addition complex
core, mineralogical and petrophysical studies may be required to establish relationships
between the log measurements and the formation properties. Typically, gamma ray,
PEF, SP and resistivity logs are used to indicate changes in formation properties as well
as the porosity logs.
The photoelectric factor, PEF, is directly related to the formation chemistry
[Equation 2.9] and can aid lithology interpretation when the volumetric components are
considered (Bateman 1985). The total volumetric cross-section, UTotal:

Equation 2.34

2-52
 Chapter 2: Porosity: Definitions, measurement and calculation

where pe = the electron density and V = the volume fraction of that material, is useful
for lithology identification (porosity must be included as one of the volumetric
components) and is a useful check for other methods of lithology identification [Figure
2.35],

Liquid-filled h o le s (p, = 1 .1 9 0 g/cm 3; C, = 2 5 0 kppm)

o Sulfur
© S a lt

O Anhydrite

<t>cNLcor- neutron porosity index (p.u.) {apparent lim eston e porosity)

Figure 2.32 Company Crossplot for density and neutron porosity estimate (CP-If
Schlumberger 1999).

2-53
 Chapter 2: Porosity: Definitions, measurement and calculation

/
t Gypsum

f tI 0 9? y
r
S e con dar/ -------
r &
Poroisity
I I
Vma = 5943 m/sec
! = 19,500 ft/sec
s / HQuartz sandstomB
M

C)alc ite (llrniBStCme]

y S' S f HV__=
1I f f ff
Cf— a o '£r rim 5486 m /sec
= 18,000 ft/sec
1 2 34
Dolomite
?
3214 1
----

cw Anlhydrite

/
. _______n 1 Su Ifur
t

Approximiate

J snaie
reninn
oFres !h mud
p , = 1 .0 Mg/rn3, •t, = 62lDpisec/ m
p, = 1 .0 g/crri3, t ,= 189Ip s ec/f t

. .. • Salt mud
pf = 1.1 Mg/m3, t , = 607 psec/m
p, = 1.1 g/cm 3, t ,=
185 psec/ft
_______ 1
-

0.3 0.4 0.5 0.6 0.7 0.8

«• Schtantanjer N

This crossplot may he used to help identify mineral mixtures

XT
from sonic, density and neutron logs. (The CNL neutron log N = -( ^ p
N )f “ ^N
— r u
( English . • ,
or metric)
is used in the above chart; the time average sonic response is
assumed.) Except in gas-bearing formations, M and N are Points for binary mixtures plot along a line connecting the
practically independent o f porosity. They are defined as: tw o mineral points. Ternary mixtures plot within the triangle
defined by the three constituent minerals. The effect o f gas,
shaliness. secondary porosity, etc., is to shift data points in the
M = tf xO.Ol (English)
Ph “ Pr directions shown by the arrows.
The dolomite and sandstone lines on Chart CP-8 are divided
by porosity range as follows: 1) <f>= 0 (tight formation);
M=— x 0.003 (metric)
Pb ~Pf 2> <|) = 0 to 12 p.u.; 3)<j> = 12 to 27 p.u.; and 4)<|> = 27 to 40 p.u.

Figure 2.33 MID plot use density, neutron and sonic logs (CP-8a Schlumberger 1999).

2-54
 Chapter 2: Porosity: Definitions, measurement and calculation

//W a te r

Conditions:
6 ma — 2.65
Qsh = 2.45
C?w = 1

Shale
Matrix

© 1969 Schlumberger

0 10 20 30 40 50 60 70 80 90 100
<t>N

Figure 2.34 Density-Neutron crossplot for shaley sandstones (Figure 8-15

Schlumberger 1989).

When only neutron and density logs are available, two different neutron and
density formulae can be used to eliminate the effect of shale (Bateman 1985):

DSh
Equation 2.35
$NSh $ DSh

^ _ $D X0NSh $N X^DSh
Equation 2.36
{^NSh ~ 0 D S h ) ~ { 0 N ~ (P d )

2-55
 Chapter 2: Porosity: Definitions, measurement and calculation

where = density porosity, <

f>
DSh = shale density porosity (known or estimated), ^ =
neutron porosity, ^NSh = shale neutron porosity (known or estimated). This empirical
method is used to offset the non-porosity effect of the two porosity measures and is
analogous with the density-neutron-gamma ray crossplot where the shale point requires
selection and suffers from the same drawbacks.
Typically, it is simple to solve the problem of shale by use of a crossplot, such
as Figure 2.34, or more commonly to use a shale indicator and an appropriate shale
property. Often the shale property (density, travel time, neutron porosity) differs from
the matrix properties in such a way that porosity is overestimated (Ellis 1987).
The shale effect can be compensated for by using the appropriate shale property
value measured in an adjacent shale bed (or estimated) and V* is estimated from the
gamma-ray log (Rider 1996)[Equation 2.33]. The appropriate formulae for density,
neutron and sonic porosity with shale are respectively:

Pm 0- Kh ) + Psh Vsh Pb
Equation 2.37
Pm P jl

Ncorrected VSh0Nsh > Equation 2.38

sh
Equation 2.39
At Jlf, —At m J

where p m = matrix density, psh = shale density, ph = bulk density, (/>scorrected = chart
corrected neutron porosity, ^ sh = shale neutron porosity, At = travel time, Atm= matrix
travel time, Atfl = fluid travel time and Atsh = shale travel time (Western 1985). The
main problem is choice of shale point that can significantly alter the porosity values.
Empirical formulae are also in significant use for sonic porosity calculation,
which are based on core. There are several formulae of the form (Khaksar et al. 1998):

_? A - B V
i _ At_________ , Equation 2.40
C

where A, B and C are empirically derived constants for core measurements. For
example, the Han equation (Han et al. 1986):

2-56
 Chapter 2: Porosity: Definitions, measurement and calculation

— -5.59 + 2.18K* _ . , ..
, _ _A/ Equation 2.41
6^93 '

This approach is often adequate to solve the problem of clay on a local or

regional level. (Khaksar and Griffiths 1998) provides a detailed comparison of sonic
%

porosity calculation methods.

Pmae VefSUS I U ,

K-Feldspar.

Calcite

Barite

Dolomite
Heavy minerals-

Anhydrite

Kaolinite

Figure 2.35 Bulk density and apparent matrix volumetric photoelectric factor (CP-21
Schlumberger 1999).

2-57
 Chapter 2: Porosity: Definitions, measurement and calculation

Note that NMR porosity is not affected to any significant extent by clay or
lithology due to the processing used to calculate porosity although pore morphology can
affect NMR porosity. New tools that have very short dead times (<5ms) can be used to
estimate the effect/volume of clay (Allen et al. 1997).

Hydrocarbons can affect log porosity estimates because of the physical properties
of the hydrocarbons. The most frequent problem is the hydrocarbon density/travel time
compared with water. The fluid density is often lower (greater travel time) than water,
especially for natural gas fields in which density and sonic measurements overestimate
the porosity, whereas neutron measurements underestimate porosity (Bateman 1985).
Generally, the correct porosity can be calculated by an appropriate choice of fluid
density [Equation 2.2] and travel time [Figure 2.36]. An exception is tar, which can be of
greater fluid density and lower travel time than water (Cunningham et al. 1992).

Figure 2.36 Fluid density estimate from density and neutron logs where Shr is the
hydrocarbon saturation and ph is the hydrocarbon density (CP-10 Schlumberger 1999).

2-58
 Chapter 2: Porosity: Definitions, measurement and calculation

The neutron porosity may be affected in a number of ways. The hydrogen index
of the formation fluids is often less than that of water, especially for gas (Ellis 1986).
This increases the neutron slowing-down length, decreasing the recorded porosity.
An additional feature of oils is the fact that carbon is a better moderator of
neutrons than oxygen (Sherman et al. 1983). This effect results in a reduction in the
slowing-down length (increased cross-section) and an apparent porosity increase
compared with water of as much as 4-5% (reverse excavation effect). The details of the
effect of crude oil are dependent on its composition compared with water. A correction
can be applied once the oil’s hydrogen index is known. This could be acquired from a
pressure tester sample.
Saturation may also need to be considered. Hydrocarbon saturation affects the
pore fluid density, hydrogen index and neutron cross-section because the neutron and
density values will be affected by the mix of the water and hydrocarbon properties
(Schlumberger 1989). This is most frequently applied when gas is present (Cuddy et al.
1994). Using Equation 2.26 if resistivity logs are available and the hydrocarbon
saturation, porosity may be corrected for hydrocarbons using a chart [Figure 2.37].
Another method uses an infrequently used empirical formulae to estimate the
correction for the neutron and density porosity (Bateman 1985) based on (Gaymard and
Poupon 1968):

0m — 2.2Ph 1.2p h, Equation 2.42

0Dh —1.7 0.7/7a , Equation 2.43

where (fDh = density porosity hydrocarbon correction, (j>Nh = neutron porosity

hydrocarbon correction and p h = hydrocarbon density. The correction is dependent on
the hydrocarbon saturation, Sw:

Equation 2.44

Equation 2.45

2-59
 Chapter 2: Porosity: Definitions, measurement and calculation

A simple and often used empirical method for calculating porosity in the
presence of hydrocarbons (especially gas) is the square root porosity:

Dcorrected t ^corrected Equation 2.46

where <
/>
Dcorrected = corrected density porosity and ^ corrected = corrected neutron porosity
(Bateman 1985). Hydrocarbon (especially gas) reduces neutron porosity, because of the
low formation hydrogen index, but increases density porosity because of the low
formation fluid density (Bateman 1985). The square root porosity does improve the
precision of the porosity estimation compared to a single porosity estimator, but does
not usually provide accurate porosity values. Tool responses and volumes of
investigation differences result in different magnitudes of the gas effect on the density
and neutron measurements (Cowan and Wright 1997).
NMR porosity is affected by hydrocarbon in a number of ways. Viscosity is
important for oils, since this increases the recorded T2 arrival time because of the
smaller surface relaxivity of oil compared with water (Kenyon 1997). Most reservoirs
are water wet and oil residing in the centre of the pores adds to the long relaxation
times. However, the signal amplitude (proportional to porosity) is usually unaffected
(but not always!). Oils containing vanadium (paramagnetic ions) have very different
surface relaxivity compared with water. This may cause the oil to relax very rapidly, so
the oil signal may be lost during the dead time before the tool starts recording the
formation response. The signal amplitude of gas is usually very low compared with
water because of the low proton concentration in the pore fluids. Therefore, low
porosity and long T2 arrival times (infrequent relaxations because of the low proton
concentration) are recorded for gas. Special processing is often required if accurate
porosity is required (Kenyon 1997).
In additional clay bound water often relaxes very rapidly and a new generation
of tools has been developed to capitalise on this property to provide an estimation of the
clay volume. Although the distribution of the bound water is dependent on the clay
structure which tends to produce underestimates of the clay volume due to bound water
not being connected directly to the main pore volume by filament like clay particles.

2-60
 Chapter 2: Porosity: Definitions, measurement and calculation

Clay is conductive and leads to a reduction in resistivity giving the impression of

greater porosity, but also more importantly masking the presence of hydrocarbons. The
effect of clay on the travel times is primarily dependent on the mechanical strength of the
shale comparatively with the mechanical strength of the matrix minerals. As a result the
shale tends to have slower travel times compared with the matrix minerals, but this is not
necessarily always the case.

The drawback with any of these combined measurement approaches to porosity

estimation is that if one of the input porosity logs is incorrect then the derived porosity
values will also be incorrect. The possibility exists for calculating a far worse porosity
value than any of the individual components.
This nomograph estimates porosity in hydrocarbon-bearing for­
mations using neutron, density and R*0 logs. The neutron and
dew 4*cor ♦l dt>cor
density logs must be corrected for environmental effects and
(CN L*) (S N P )
lithology prior to entry into the nomograph. The chart includes
_ so _ r 50 50 _ an approximate correction for excavation effect, but if ph < 0.25
(gases), the chart may not be accurate in some extreme cases:
! (pu.)
very high values of porosity (> 35 p.u.) coupled with medium to
high values o f Shi, and for Shi - 100% for medium to high values
o f porosity.
To use, connect the apparent neutron porosity point on the
appropriate neutron stem with the apparent density porosity on
the density stem with a straight line. From the intersection o f this
1 40 2 1 40 1 40 J 40 line with the porosity, <pi, stem, draw a line to the origin o f the
Shr versus Ad chart. Entering this chart with the hydrocarbon
saturation, Shr, (Shr ” 1 - Si0) defines a porosity correction factor
A<fi. This correction factor algebraically added to porosity, di,
gives the true porosity.

Example: dcNLeor” 12 p.u. giving di - 32.2 p.u.

< j> C )c o r= 3 8 p .U . and A d - -1 .6 p.u.

L 30 J L 30 Shr-5 0 % Therefore, d - 32.2 - 1.6

1 30 J 30
- 30.6 p.u.

_ 20 2 _ 20 L 20 J 20

Ad
(p.u.)
L 10 J 1 10 L 10 J 10

-1

_ 0 _ 0 _0
100 80 60 40 20 0

S h ,(% )

Figure 2.37 Porosity correction hydrocarbon (CP-9 Schlumberger 1999).

2-61
 Chapter 2: Porosity: Definitions, measurement and calculation

2.4.2 Numerical solutions

A straightforward approach to porosity estimation may not be valid for example,

complex lithologies in which the matrix of the formation may be limestone with
significant dolomite and clay, a shaly arkose sandstone, or the presence of heavy
minerals (Bateman, 1985; Doveton, 1994; Ellis, 1987; Rider, 1986; Schlumberger,
1989). Often, if more than one porosity tool is available, the different porosity estimates
can be used to check or correct the porosity values calculated. However, this may still
provide inadequate porosity estimates when compared with core porosity values due to
the clay contents.

Empirical relationships may be used in these cases, but are often field or region
specific. A number of techniques are used to derive porosity (and other) relationships:
polynomial regression, mineral inversion, factor/principal-component analysis, cluster
analysis, neural networks, genetic algorithms, and fuzzy logic (Doveton 1994; Elphick
et al. 1996; Cuddy 1997). These methods all normally require the use of additional log
values. The resistivity logs such as SP, laterolog, induction and electromagnetic
propagation are less sensitive to lithology than the other logs and so are not always
useful for differentiating complex lithologies, but nevertheless can resolve complex
cases (Pickett 1973).
The problem with some of these techniques (neural networks and genetic
algorithms) is that they may solve the problem of calculating porosity, but they may not
provide any addition information. In other words, it is often difficult to relate any
physical reasoning to the formulated relationship, for example the presence of micas.
Another method of obtaining accurate porosity in complex lithologies and facies
is to run advanced logging tools (Lofts et al. 1995). The geochemical tool (Elemental
Capture Spectroscopy tool) can provide the abundance of numerous elements (Si, Al,
Ti, Fe, Ca, K, S, Gd, Th, U, indirectly Mg + Na) from which the formation mineralogy
may be calculated. From a detailed mineralogy the formation density, and therefore
porosity, may be calculated with increased accuracy (compared with density tools).
However, the geochemical tool is very rarely used for oilfield evaluation because of the

2-62
 Chapter 2: Porosity: Definitions, measurement and calculation

The Gassmann (1951) equation may be used to improve sonic porosity

estimates. The Gassmann equation considers the mechanical properties of the formation
and the fluids. The formation mechanical properties are dependent on the formation
bulk and shear rigidity. The fluid properties are dependent on fluid salinity, dissolved
gas, fluid pressure, temperature, water saturation, gas/oil ratio and fluid density. These
properties may be described by the bulk moduli of the dry rock frame, the rock grains
and fluid, and the shear modulus of the dry rock frame. The Gassmann equation
(Gassmann 1951) re-arranged for porosity is:

+ - 1 , Equation 2.47
A _

where Kfr - dry frame bulk modulus (dynes/cm2), Gfr = dry frame shear modulus
(dynes/cm2), Kgr = grain bulk modulus (dynes/cm2), Kfl = fluid bulk modulus
(dynes/cm2), pb = bulk density (g/cm3) and At = compressional wave travel time
(m/psec).
The bulk moduli of the dry rock frame, the rock grains and fluids may be
measured in the laboratory or are known constants (Alberty 1996):

Equation 2.48

Equation 2.49

where Kx = bulk modulus of the appropriate material, Gx = shear modulus of the

appropriate material, Vpx = compressional velocity of the appropriate material, Vsx =
shear velocity of the appropriate material and px = density of the appropriate material.
Several investigations of the magnitudes of altering the different parameters and the
consequences (Whitman and Towle 1992; Holbrook et al. 1999). Other formulation are
used such as Biot-Stoll model which is examined in Leurer (1997) and probably the
most frequently used variant of the Gassman model (Xu and White 1995).

2-63
 Chapter 2: Porosity: Definitions, measurement and calculation

When an array or dipole sonic tool is used in conjunction with a density tool the
compressional and shear wave velocities and bulk density are measured continuously.
Thus the dry frame bulk and shear moduli may be inferred and the bulk moduli for
grains are published in data books (Alberty 1996). The bulk moduli for brines may be
calculated using a chart given the salinity, temperature and pressure. However, the
values of K and G should always be measured on core samples since textbook values
are likely to be inaccurate. The effects of hydrocarbon can be significant. However,
many studies exist on these effects and, with the aid of modelling the appropriate
moduli values can be derived (Bourbie et al. 1987; Alberty 1996) and require significant
core data to calibrate the equations.

2-64
 Chapter 2: Porosity: Definitions, measurement and calculation

2.4.3 Porosity: Approach used for thesis

This section describes techniques that are used in the data chapter (chapter 4) of
this thesis (Gaymard and Poupon 1968; Wiley and Patchett 1994; Cowan and Wright
1997).

Gaymard

Gaymard and Poupon (1968) considered the effects of hydrocarbon and mud
salinity on density and neutron measurements. A theoretical approach for log correction
was derived, given that porosity may be more accurately calculated when the density,
neutron absorption and hydrogen index of the formation fluids are accounted for. The
theory assumes that the neutron and density tools measure the porosity of the invaded
zone, which is saturated with mud filtrate and residual hydrocarbons with hydrocarbon
density between 0 and 0.9 g/cm3. The formulae derived for density and neutron porosity
estimates are (Gaymard and Poupon 1968):

[ft(l.674+3.24/>, + 9p i )]-l + 0.4P

0N =<f 1+ S t Equation 2.50
1-0.4P

’ 1.07St [l.ll + 0 .6 5 P -A (l.l86 + 0.36p h + p \ f

Equation 2.51
p a -1 -0 .7 P
r'm

where (j>D= density porosity, ^ = neutron porosity, (/= true porosity, Sh = hydrocarbon
saturation, ph = hydrocarbon density and P = mud filtrate salinity. This method includes
the presence of residual formation water. It ignores the possibility of significantly
different salinity compared with the mud filtrate. In addition, through gas-oil transition
zones, gas, oil, formation water and mud filtrate will all be present. In such cases the
above formulae are unable to realistically represent the combinations of formation
fluids. However, this method should still improve porosity estimates if the
estimates/measurements of the mud filtrate salinity, hydrocarbon density and
hydrocarbon saturation are accurate as would be available for a mature oilfield. Note:
the effects of hydrocarbon will be greater at large formation porosity values. Porosity

2-65
 Chapter 2: Porosity: Definitions, measurement and calculation

estimates from the density and neutron tools maybe very different if invasion is less
than lOinches (25cm) because of their differing depths of investigation (Cowan and
Wright 1997). This method is rarely used as a simpler empirical method for calculating
porosity will have been established prior to all the required information being available.

Wiley & Patchett

Wiley and Patchett (1994) used several different numerical codes to derive
formulae for density and neutron measurements to calculate porosity estimates. The
effects of lithology, invasion, tool investigation depths and reservoir fluids were
considered. The codes used for the neutron and density measurements were three-
group, three-dimensional, three region diffusion codes, which were benchmarked
against stochastic Monte Carlo codes.
Accurate porosity (±lpu) can be problematic; especially in high porosity, gas
saturated reservoirs due to the uncertainties in the responses of the porosity tools in
these circumstances. The formulae were developed primarily due to the inaccuracies of
the standard square root porosity estimator for gas reservoirs [Equation 2.46]. The
(Wiley and Patchett 1994) porosity formula derived for clean (shale free) sandstone is:

^ = 1.916 + 0.4253^, +1.828E~30 2N -1.885£"5^ -0 .4 9 0 3 ^

, , , , , Equation 2.52
+ 3.SS2E ^ -0.4397/s ( ^ / 100)

where </D= density porosity (matrix density = 2.65 g/cm3 and fluid density = 1 g/cm3),
(f>N- neutron porosity (sandstone scale). Additional formulae were also derived for
alternative lithologies. This method is not in common use due to the modelling
overheads required for the specific formation and borehole conditions encountered;
there is significant appeal for generating porosity values given the correct input data.

2-66
 Chapter 2: Porosity: Definitions, measurement and calculation

Cowan & Wright

Cowan and Wright (1997) studied anomalous porosity values in horizontal gas
wells. Details of porosity estimation methods were given including the square method
[Equation 2.46], (Wiley and Patchett 1994) [Equation 2.52] and a shallow iterative
method. The shallow iterative method consisted of using:

1) The square root porosity with chart corrected neutron porosity values and density
porosity assuming a sandstone matrix.
2) Calculating the hydrocarbon saturation (1-SW) using Archie’s equation [Equation
2.26] with the square root porosity.
3) Estimate the hydrocarbon density using a chart [Figure 2.36] or assume it is known.
4) Calculating corrections for bulk density and neutron porosity using the (modified)
Gaymard and Poupon (1968) equations given below.
5) Add the corrections and calculate new density and neutron porosities.
6) Repeat until convergence.

Cowan and Wright (1997) modified Gaymard equations (Gaymard and Poupon
1968):

APb = ^ K 1-19 -0 .1 6 P )p m/ - A p h - R j , Equation 2.53

\(l -P )P m f-C p h ~£>\

A0 n ~(j)Sh------------ 7-— -r------- + A(/fNex, Equation 2.54
PmfV-P)

where fa = density porosity, fa = neutron porosity, </)= true porosity, Sh = hydrocarbon

saturation, p mf = mud filtrate density, ph = hydrocarbon density, P = mud filtrate salinity
and Afaex = excavation correction. For oil reservoirs, A = 1.19, B = 0.032, C = 1 and D
= 0.3 and for gas reservoirs, A = 1.33, B = 0, C = 2.2 and D = 0.
The shallow iterative method provides water saturation values. However, it
includes the same pitfalls as the Gaymard and Poupon (1968) method, but are largely
compensated for by the process of iteration. Cowan and Wright (1997) suggested
replacing the square root equation with the Wiley and Patchett (1994) equation in the
shallow iterative method would further improve the accuracy of the porosity estimation.

2-67
 Chapter 2: Porosity: Definitions, measurement and calculation

2.5 Summary

This chapter defines porosity in a geological and petrophysical context. An

explanation of the basis of porosity estimation from core samples and openhole wireline
and LWD logs is given. The theory and methodology covers density, neutron, sonic,
NMR and resistivity tools and includes several core-based porosity measurements. The
limitations of these techniques are explained and techniques to improve porosity
estimation from log measurements are detailed with reference to clay.
Gas expansion core porosity is taken to be representative of the formation
porosity for the purposes of this thesis. It is stressed that no method (core or logging)
directly measures porosity. Porosity can only be estimated based on a related formation
property. Density porosity estimates are the most reliable log porosity estimator in most
circumstances. However, the use of more than one porosity estimation can improve
accuracy. The agreement of more than one porosity estimation technique adds to the
confidence in the porosity values estimated, but does not provide proof of correct
porosity values. Simple techniques such as crossploting can improve porosity estimates.
The gamma ray log can be used to estimate the volume of clay/shale that in turn
can be accounted for. However, care must be taken because gamma ray logs can be
affected by many factors related and not related to the presence of clay minerals.
Likewise the photoelectric factor can be used to account for mixed lithologies, but
maybe affected by weighted muds.
Specialist techniques such as polynomial regression, mineral inversion,
factor/principal-component analysis, cluster analysis, neural networks, genetic
algorithms, and fuzzy logic are mentioned and can be used in complex situations.
However, these methods are time consuming and may not improve porosity estimation.
The descriptions given above rely heavily on published Schlumberger
documents. The fundamentals of the measurements have firm groundings in physics and
vary little between Service Companies. The main differences between service
companies occur in the data collection and processing and not tool design where
detectors and the spacings are all fundamentally the same wireline and LWD tools. Note
that more than half the data used in this thesis was recorded by Schlumberger.
Tool related effects

2-68
 Chapter 2: Porosity: Definitions, measurement and calculation

• The effects of non-parallel standoff can result in poor porosity values being
recorded and often little or no indication of a problem is evident from the logs,
although analysis of the individual detector count rates can provide an indication.
For sonic tools, non-parallel standoff is accounted for in the tool design. There
seems little recognition in the general literature of these(often significant) effects.
• Toolstring dressing (bowspring, cranks...) is important, since imbalance of the
mechanical forces acting on the entire toolstring can impart turning motions leading
to non-parallel standoff.
• Mud balling/rings (build up of mud on the tools e.g. on the leading edge of the
density pad) can cause turning motions and non-parallel standoff.
• An under-gauge caliper log may indicate mudcake, but could be the tool running
across a chord rather than the diameter of the borehole giving non-parallel standoff.

Clay/Shales

• Log responses in shales can be markedly different and can cause problems when
making shale corrections to logs in shaly sand, often resulting from shales being out
of the calibration range of the tool. Shale may contain a variety of rare earth
minerals in minute quantities and variable amounts of bound water that can affect
log responses.
• Problems exist with non-radioactive shales, radioactive sands, micas, feldspars,
potassium evaporites and U rich formation fluids, U-salt deposits in organic matter
or hardgrounds (Ellis 1987; Hurst 1990; Rider 1996).
• The effect of clay on the sonic travel time is primarily dependent on the mechanical
strength of the shale compared with the mechanical strength of the matrix minerals.
Shale tends to have slower travel times compared with the matrix minerals, but this
is not necessarily always the case.
• Clay distribution within the formation alters the choice of shale volume equation as
this impacts the volume proportion of the formation calculated that is clay and
therefore the porosity values (Katahara 1995).

2-69
 Chapter 2: Porosity: Definitions, measurement and calculation

• No logging tool provides an accurate measure of clay volume or distribution,

therefore many compromises have to be made. Until an accurate clay-logging tool
is developed, clay estimation will remain an art not a science.
• Specific relations between the measurable properties of minerals in a core lab are
particularly difficult e.g. X-ray diffraction mineral identification may be in error by
50% or more (Doveton 1994). These errors are amplified in the logging
environment.
• The principal drawback is that the choice of shale/clay point is subjective (Society
2001), and therefore implementation is rarely repeatable.

Noise

• Sonic log quality is mostly affected by noise sources of one type or another
particularly from the actions of tool jewellery against the borehole. Stacking the
waveforms from each receiver provides some immunity for array and dipole tools.
The rejection of poor data from receivers improves the data (Kimball and Scheibner
1998).
• There may be operational reasons for correction factors, such as the use of different
companies’ logging tools and different generations of tools or due to tool minor
malfunctions (e.g. poor calibration).
• Additional problems in rugose and enlarged boreholes are often cyclic and require
careful filtering (Nieto et al. 1995).

Hydrocarbon effects

• Hydrocarbons can affect log porosity estimates. The fluid density is often lower
(greater travel time) than water, especially for natural gas fields in which density
and sonic measurements overestimate the porosity, whereas neutron measurements
underestimate porosity (Bateman 1985). The hydrogen index and capture neutron
cross-section of the formation fluids is often less than that of water, especially for
gas (Ellis 1986). This increases the neutron slowing-down length, decreasing the
porosity (Sherman et al. 1983).

2-70
 Chapter 2: Porosity: Definitions, measurement and calculation

• The effects of hydrocarbon can be significant on elastic moduli values of

formations and can alter sonic porosity values.

Multi-log porosity estimation

• The drawback with any of these combined measurement approaches to porosity

estimation is that if one of the input porosity logs is incorrect then the derived
porosity values will also be incorrect. The possibility exists to calculate a far worse
porosity value than any of the individual components.
• Accurate porosity in complex lithologies and facies may be achieved using data
from advanced logging tools (Lofts et al. 1995), but because of the tools, poor
performance in resolving different minerals and the non-unique chemistry of rock
forming minerals (Harvey et al. 1992) can give poor results.
• Elastic moduli values require significant core data to calibrate the equations which
may not be available.
• Tool responses and differences in volumes of investigation result in different
magnitudes of the gas effect on the density and neutron measurements for example
(Cowan and Wright 1997).
• Core data is often very sparse and therefore a core-log relationship may not
necessarily be applicable on field wide basis.

Log quality

An extensive set of papers on log quality that deals with most of the issues
raised in this chapter can be found in The Log Analyst (LPS 1994). The papers describe
the problems of all core, wireline and LWD data collection and the potential effects on
interpretation. In addition, Myers (1992) provides a thorough review of neutron logging
and Brami (1991) provides a discussion of the effects of the different LWD tools
offered by several Service Companies.

2-71
 Chapter 3: Horizontal well porosity anomalies

Chapter 3: Horizontal well porosity anomalies

This chapter reviews publications to provide an overview of present knowledge

of horizontal well porosity anomalies. Firstly, horizontal wells and their uses are
explained. Secondly, observed porosity anomalies in horizontal wells are described and
the potential causes are detailed.

3.1 Horizontal wells and their application

Definitions of what constitutes a horizontal well vary considerably. “A

horizontal well is a well drilled approximately horizontally within the confines of a
reservoir approximately parallel to the formation boundaries” (Fritz 1991). For wireline
logging tools any well greater than 60° (angles are always measured from vertical) will
require pipe or coiled tubing conveyed logging to push the tools down the borehole.
However, the term ‘horizontal well’ is used more frequently to refer to any well that
deviates significantly from vertical (>75°). The diameter, length and radius of build
(rate at which the angle of the borehole inclination is increased) of horizontal wells vary
considerably depending on their purpose and are classified by radius of build [Figure
3.1](Bigelow et al. 1992). The radius of build of the borehole controls whether or not
logs can be run and/or core can be extracted [Figure 3.1].
In recent years increasing numbers of horizontal wells have been drilled to
improve reservoir drainage (Weber 1999). Drilling horizontal wells offshore is
expensive and must be justified as the most cost-effective manner of overcoming a
specific production problem, or simply the most cost efficient way to drain the
reservoir. The production problems that may be encountered are:

• Tight formations. Horizontal wells open more migration paths or are possibly used
as injector wells to provide water/gas drive or to induce fractures into the formation.
• Fractured formations. Drilling perpendicular to the fracture orientation means as
many fractures as possible are intersected, often this requires horizontal drilling
because fractures are oriented in the plane of minimum stress (often vertical).
• Channel sands. Drilling horizontally to intersect as many channels as possible.

3-1
 Chapter 3: Horizontal well porosity anomalies

Water and/or gas coning problems can be delayed by using horizontal wells.
Permeability barriers. Horizontal wells can drain otherwise inaccessible parts of the
reservoir.

Borehole

Short Radius - (20-40ft)

Typically, no logs or cores
Reach <650 ft (* 2 0 0 m)
(A) Build 1.5°-3°/ft

Medium Radius -(300-700ft)

Usually limited to small
diameter, flexible tools
Reach < 2.300 ft (* 700 m)
(B) Build 8°-207100ft

Long Radius - (1000-3000ft)

Can usually accommodate
core barrel and most
wireline devices
Reach > 3.000 ft (* 900 m)
Build 2°-67100ft
(C)

Figure 3.1 Schematic to demonstrate different types of horizontal wells (Figure 9

Bigelow et al. 1992) with radius and build information (Figure 3 Fritz 1991).

3-2
 Chapter 3: Horizontal well porosity anomalies

3.2 Observed porosity anomalies in horizontal wells

Examples of porosity anomalies in horizontal well logs and different logging

techniques are presented below. Possible sources of the perturbations and their likely
affect on the recorded porosity values are discussed.
In recent years the validity of log-derived porosity and water saturation
estimates from horizontal wells has been questioned. Numerous authors have suggested
possible causes for high porosity values in horizontal wells (Allen et al. 1990; Betts et
al. 1990; Day and Petler 1990; Woodhouse et al. 1991; Bigelow et al. 1992; Austin et
al. 1994; Cuddy et al. 1994; Bedford et al. 1997; Cowan and Wright 1997; Ellis and
Chiaramonte 2000). However, no comprehensive solutions have yet been put forward
that explain convincingly the physical processes that result in anomalous porosity
values from horizontal wells. The challenge is to produce robust guidelines for the
interpretation of horizontal well porosity logs. Unfortunately, cores are seldom
recovered from horizontal wells and direct comparison with the vertical well core
porosity values may indicate anomalous porosity values in the horizontal well, but the
cause of the porosity anomalies maybe ambiguous. Great care must be taken when
comparing vertical well measurements with horizontal well measurements since
lithology and facies, borehole and formation conditions may be radically different.
Logging horizontal wells is often problematic for operational reasons, even when
conditions are favourable.
A number of problems may adversely affect the log measurements from
horizontal wells, which do not affect vertical well logs to the same degree for example,
anisotropic mud invasion (either by gravity or high permeability anisotropy [K h /K v] )

(Woodhouse et al. 1991; Cuddy et al. 1994), debris (Cuddy et al. 1994),
fractured/unstable borehole (Allen et al. 1990; Austin et al. 1994; Cuddy et al. 1994),
bed boundary dip (White 1991), oval/enlarged hole (Cuddy et al. 1994) and preferential
tool orientation (White 1991; Cuddy et al. 1994) can all affect the accuracy of porosity
estimations from horizontal wells [Figure 3.2].
Note that Cuddy et al. 1994’s conclusions with regard to the incorrect pad
alignment scenario [Figure 3.2] are not bom out by experience of in excess of 100
horizontal wells analysed. In addition from observation of the diagram in Figure 3.2, it
is clear that the caliper will read under bit size where the borehole is not washed out. It

3-3
 Chapter 3: Horizontal well porosity anomalies

is rare to observe an effect on the density correction log (DRHO), but not unknown and
is dependent on mud type as DRHO compensates for parallel standoff. In this case the
standoff is non-parallel therefore DRHO does not provide good standoff correction to
bulk density, RHOB.

Formation

Asymmetric invasion.
Invading fluid ( Borehole
This invasion profile is produced by asymmetric invasion
because of high Kv/Kh ratio associated with the aeolian deposits.
Typical Kv/Kh values lie in the range 5-10.
Density measures here

“Formation relaxation
fractures”
Fracturing.
Stress deformation of wellborc may increase pore space.
However little or no sign of washouts were seen that would
fractures support such an idea, and any fractures would normally be
seen at the side, not the bottom.

Mudcake

Borehole Borehole

Invading fluids drop

away from borehole.

♦
Keyseat cut in borehole, possibly
Filtrate
re-filled with cuttings.

Hole condition affected by keyseat

Loss of filtrate by gravity segregation.
A groove cut in the low side of the weilbore might give
It is possible that the filtrate could fall down and away a low density, particularly if filled with cuttings.
from the weilbore due to gravity segregation. However no effects on DRHO or caliper logs were seen
However this would require high vertical permeability, which would be expected.
which is not observed in these wells.

Incorrect pad alignment

Tool lies on its side
measuring mud as
well as formation. Borehole Incorrectly oriented drillpipe may align the density tool sideways
rather than down. This poor pad application might read tool low
density.
However, no effects on DRHO or the caliper were seen that would
support this.

Figure 3.2 Horizontal well logging environment and the effect on measurements
(Figure 3 Bedford et al. 1997).

3-4
 Chapter 3: Horizontal well porosity anomalies

In the North Sea new vertical wells are often logged with wireline and/or LWD
tools, but horizontal wells are generally only logged with LWD tools (Rider 1996). The
differences between wireline and LWD tool responses can be significant (Coope 1983;
Sakurai et al. 1992) as well as perturbations due to tool to formation contact, formation
conditions, well conditions and the lithologies encountered.

• Wireline and LWD nuclear tools are not calibrated in shales therefore their
readings will differ from one another and differ between Service Companies.

The above perturbations will generally lead to increased porosity measured in

horizontal wells, as has been demonstrated by BP in four North Sea fields [Figure
3.3](Bedford et al. 1997). Figure 3.3 demonstrates that when considering horizontal
porosity anomalies wells >75° inclination should be considered as horizontal since
significantly lower densities (higher porosities) are observed at these deviations and
will be used throughout this thesis. However, an exception is logging while drilling
(LWD) neutron porosity values, which are expected to be slightly lower in a horizontal
well than those observed in a vertical well, because of the eccentring of the LWD tool
(Day and Petler 1990). For resistivity tools the main effect in horizontal wells is the low
borehole to bed angle, which requires modelling and has been well documented
(Anderson et al. 1990; Anderson et al. 1992; Gianzero et al. 1992; Vernon et al. 1993;
Li et al. 1994; Anderson et al. 1996; Koelman et al. 1996; Wu et al. 1997). In contrast,
there have been relatively few investigations that have included modelling of porosity
tool responses (Day and Petler 1990; Austin et al. 1994; Cuddy et al. 1994; Bedford et
al. 1997; Cowan and Wright 1997; Ellis and Chiaramonte 2000; Cannon and Kienitz
1999).

One long-term study described a large number of vertical and horizontal gas
wells from the southern North Sea that were examined for porosity anomalies (Cuddy et
al., Day et al. 1994; Bedford et al. 1997; Cowan and Wright 1997). The study illustrated
that log porosity values overestimated porosity in horizontal wells regardless of tool
type [Figure 3.3 and Table 3.1]. The overestimates appeared to increase with increased
porosity and permeability.

3-5
 Chapter 3: Horizontal well porosity anomalies

Gas Average Porosity Average

Number of Weils
Field (Horizontal - Vertical) (pu) Porosity Kfc/Kv Tool Types
Vertical Horizontal Density Neutron Sonic (pu)
A 25 3 6 -2 2 11 7.1 Wireline
B 18 2 4 1 2 13 5.2 Wireline/LWD
C 17 1 9 4 12 16 9.8 Wireline
D 11 3 15 2 N/A 11 4.9 Wireline/LWD

Table 3.1 Horizontal/Vertical well porosity comparison (Table 1 Cowan and Wright
1997).

Field A
1.900
2.000
2.100 I
8 2.200 1 1t7
2.300 It -

1 F™ II
2-400
2.500
fc h d
t----------
KT — i------------------ 1----------------- -------- 1---------
2.600

HoleAngle (Degrees)
Field B
2.1501
2.200
2.250
2.300
2.350
2.400
2.450
2.500 :
2.550*1
2.600
2.650
40 60
HoleAngle (Degrees)

Field C
2.150
2.200
2.250
S 2.300 r j
g 2.350
£■ 2.400
* r-r
f 2.450
I 2.500' 1 -------- ^
2.550
2.600
2.650 ---------- 1-----------1------ — 1 ■1-----
20 40 60 80
HoleAngle (Degrees)

Field D

■ - f a ------------ -—- r — i p ^ '

1A r j —

20 40 60 80
HoleAngle (Degrees)

Figure 3.3 Comparison of density logs with borehole inclination from four fields with
measurements from more than 80 wells (Figure 2 Bedford et al. 1997).

3-6
 Chapter 3: Horizontal well porosity anomalies

The main source of perturbation was thought to be the well-known permeability

anisotropy of the southern North Sea Rotliegend formation (Bedford et al. 1997). In
horizontal wells the invasion will be generally greater at the sides of the borehole than
below it because of permeability anisotropy (Woodhouse et al. 1991; Bedford et al.
1997; Cowan and Wright 1997). A porosity tool lying on the floor of a horizontal
borehole would measure lower porosity (apparently more gas) in this situation, in the
absence of fractures at the sides of the borehole, compared with the same tool in a
vertical well (Bedford et al. 1997).
The author’s experience of Canadian horizontal heavy oil wells where
permeability anisotropy is insignificant, gas is ‘seen’ above the borehole roof even
though nuclear tools are almost insensitive to the region behind them. If gas detection is
a priority running the nuclear tools along the well roof should be considered. Tailoring
the wireline toolstring configuration and/or processing algorithms (for LWD also) in a
non-standard manner can meet the objectives of the logging program.
One horizontal well was studied using a wireline toolstring that included two
density tools orthogonal to one another; included in the toolstring were sonic, resistivity
and ultrasonic imaging tools (Bedford et al. 1997). The results demonstrated that, in this
case, the downward facing wireline density tool porosity values measured higher
porosity compared with vertical wells in that particular field [Figure 3.4]. The side
facing density porosity values were expected to measure the true formation porosity.
However, unexpectedly the side facing density measured greater porosity values than
the downward facing density tool. The ultrasonic images revealed that there were
fractures in the side of the borehole, which increased the formation porosity calculated
[Figure 3.35]. The sonic porosity values were comparable with the vertical well values.
The solution was to gas correct the density porosity values using the calculated water
saturation values. A good match with the sonic porosity values could then be achieved.

• Running two orthogonal density tools as in the case of (Cuddy et al. 1994) to
attempt to obtain at least one good quality density log can lead to the two density
tools pulling towards one another giving two poor density logs (Samworth 2000).

3-7
 Chapter 3: Horizontal well porosity anomalies

Sonic porosity Density porosity Density porosity Side looking density

from standard from standard field assuming (1 -Sw) porosity from standard
field transform. transform. is gas. field transform..

S ide SfD 1

Sonic Down density Down density Side density

Borehole

V ® / “

Figure 3.4 Schematic of results to demonstrate improvement of horizontal well porosity

estimation from (Figure 5 Bedford et al. 1997).

The study concluded that wireline sonic porosity was generally providing more
reliable porosity values in horizontal gas wells (Note: Not always the case, see Austin et
al. 1994 below). However, borehole conditions (fractures) and the proximity of ‘fast’
beds may affect the sonic porosity values, in which case the density porosity values
would be used.
In the few horizontal wells in which sonic tools are run (approximately 5 to
10%), the sonic porosity seldom matches the other calculated porosities (Samworth
2000). This is especially so in low porosity formations (<5pu) in which the sonic travel
time appears to be independent of porosity, although this is not bom out by Austin et al.
(1994) in which porosities were >5pu (see below). Sonic tools are rarely run in
horizontal wells due to the sonic tool’s physical weakness and thus the increased risk of
toolstring loss.

3-8
 Chapter 3: Horizontal well porosity anomalies

The permeability anisotropy resulted in shallow invasion below the borehole

(<3cm) where the density readings were made and high porosity values were estimated
due to poor fluid density estimation due to the presence of gas. Modelling was
commissioned to support these results (Cowan and Wright 1997). The porosity
estimation could be resolved by using wireline shallow resistivity or epithermal neutron
tools to measure the gas saturation or if LWD tools were run to use the azimuthal
density neutron tools available (Bedford et al. 1997). Several recommendations were
made (Cuddy et al. 1994):

• Run an LWD azimuth density and neutron porosity tool.

• Shallow LWD resistivity needed to measure Sx0 and Sg.
• Time lapse LWD and wireline measurements allow time lapse studies of dynamic
changes.
• Shear wave velocity values may solve porosity measurement problems because the
shear wave is insensitive to fluid.
• Horizontal core can be used to calibrate porosity values.

The first two recommendations have subsequently been examined by Cannon

and Kienitz 1999 in a deviated well (see below).

• Time lapse density porosity from up and down wireline logs in which several hours
have elapsed between passes do not appear to show any discernible differences from
some 100+ horizontal wells.

A more detailed study has investigated southern North Sea acoustic log
responses; core acoustic travel times and core porosity values demonstrated
discrepancies between the expected and observed sonic porosity values (Austin et al.
1994). The observed sonic travel time values were 2-3ps/ft lower than expected at 6pu
porosity, but approximately lOps/ft lower than expected at 15pu porosity. This would
have led to overestimates of porosity by 5pu and underestimates of rock strength,
resulting in conservative well completion strategies (Austin et al. 1994). Conventional
core analysis, travel time readings in nearby calibration beds (5Ops/ft anhydrite), core
acoustic measurements [Table 3.2 and Table 3.3] and porosity values for the formation

3-9
 Chapter 3: Horizontal well porosity anomalies

from nearby vertical wells [Table 3.4] suggested that the sonic travel time logs were
anomalous and not the other porosity logs. Image logs were run to assess situations
where ‘fast’ beds were influencing the sonic logs. The acoustic core measurements and
theoretical calculations were unable to account for the anomalous sonic porosity values,
either by the presence of gas (as a result of anisotropic invasion) or in-situ stresses
around the borehole [Table 3.3](Austin et al. 1994). Note that these observations
contradict Cuddy et al. (1994) and Bedford et al. (1997)’s results above.

Sample Wireline Log Travel Time Difference Log-Core

Porosity (pu) Core Travel Time (ps/ft)
Number (ps/ft) Travel Time (ps/ft)

1 4.3 64.4 62.5 -1.9

2 7.1 78.1 76.3 -1.8
3 11.8 75.7 74.5 -1.2
4 11.1 74.5 73.3 -1.2
5 10.5 72.1 71.0 -1.1
6 12.9 73.9 73.7 -0.2

Table 3.2 Comparison of core and log travel time values in vertical wells (Table la
Austin etal. 1994).

Sample Porosity Core Travel Wireline Log Travel Difference Log-Core Theoretical Difference due to
Number (pu) Time (ps/ft) Time (ps/ft) Travel Time (ps/ft) Gas Saturation* (ps/ft)
1 17.8 84.4 94.8 10.4 6.7
2 19.3 84.5 91.5 7.0 7.1
3 18.1 82.7 92.3 9.6 6.1
4 17.1 80.2 90.3 10.1 5.4
5 16.5 80.4 90.6 10.2 6.0

* assumes 40su gas saturation, 60su iquid saturation.

Table 3.3 Comparison of core and log travel time values in horizontal wells (Table lb
Austin et al. 1994).

Average Density Average Sonic Average Vertical Density Derived Sonic Derived Porosity
Horizontal
Derived Porosity Derived Well ‘Target’ Porosity minus minus ‘Target’
Well
(pu) Porosity (pu) Porosity (pu) ‘Target’ Porosity (pu) Porosity (pu)
A 13.5 18.4 14.4 -0.9 4.0
B 16.7 22.4 14.2 2.5 8.2
C 12.7 Sonic Failed 13.1 -0.4 N/A
D 11.2 15.2 10.8 0.4 4.4

Table 3.4 Comparison of log sonic and density porosity values in horizontal and
vertical wells (Table 2 Austin et al. 1994).

3-10
 Chapter 3: Horizontal well porosity anomalies

Recommendations for sonic logging in horizontal wells were (Austin et al.

1994):

• Sonic velocity values must be calibrated with core.

• Log responses can often be explained by borehole images.
• Sonic velocity values are slow in horizontal gas wells because the sonic travel time
log values are too high in high porosities.
• Core sonic porosity match non-sonic log porosity.
• Sonic logs cannot be used for porosity or rock strength calculations in horizontal
gas wells.
• Shear wave velocity values may provide a solution because the shear wave is
insensitive to fluids.

Summary o f observed porosity anomalies

Anomalous porosity values are observed in horizontal wells when compared

with vertical wells within the same reservoir unit. Two studies (Cuddy et al. 1994;
Austin et al. 1994) from the North Sea both concluded that horizontal well porosity
values were too high compared with expected values. A number of recommendations
(below) were made some of which are examined in the next section of this chapter.

• Run an LWD azimuth density and neutron porosity tool to search azimuthal
porosity values for the most representative porosity value (Cuddy et al. 1994).
• Shallow LWD resistivity needed to measure Sx0 and Sg for hydrocarbon corrections
to porosity logs (Cuddy et al. 1994).
• Time lapse LWD and wireline measurements allow time lapse studies of dynamic
changes and potential a measure of formation permeability (Cuddy et al. 1994).
• Shear wave velocity values may solve porosity measurement problems because the
shear wave velocity is insensitive to fluid (Cuddy et al. 1994; Austin et al. 1994).
• Horizontal core can be used to calibrate porosity values. (Austin et al. 1994).
• Log responses can often be explained by borehole images allowing the rejection of
data affected by fast beds (Austin et al. 1994).

3-11
 Chapter 3: Horizontal well porosity anomalies

• Sonic velocity values are slow in horizontal gas wells because the sonic travel time
log values are too high in high porosities (Austin et al. 1994).
• Sonic logs cannot be used for porosity or rock strength calculations in horizontal
gas wells (Austin et al. 1994).

The author’s opinion is that the problems of logging tool deployment (either
LWD or wireline) in horizontal wells is not sufficiently addressed in published papers.
The author’s experience is that wireline tools may not be running in-hole as intended
and that inclination and deviation data should routinely be recorded for all horizontal
wells (including LWD) as log quality control. Common calibration standards for
wireline and LWD tools in limestone, dolomite, sandstone and importantly shale need
to be established, although the CALLISTO facility does address this need for limestone,
dolomite and sandstone (Samworth and Lovell 2000). Additional observations are:

• Time lapse density porosity from up and down wireline logs in which several hours
have elapsed between passes do not appear to show any discernible differences.
• Wireline and LWD nuclear tools are not calibrated in shales therefore their readings
will differ from one another and differ between Service Companies. This affects
shale volume corrections to porosity logs.
• Running two orthogonal density tools can lead to the two density tools pulling
towards one another, resulting in poor alignment of both tools against the borehole
and giving two poor density logs (Samworth 2000).

3-12
 Chapter 3: Horizontal well porosity anomalies

3.3 Perturbation sources

The sources of porosity and tool response anomalies can be broken down into
three categories, 3.3.1 tool errors, 3.3.2 borehole conditions and 3.3.3 formation
conditions. The questions are:

• Is a tool response due to a formation property, borehole condition or a tool

measurement error?
• What methods are the best for estimating the actual formation properties?

The problem is less straightforward than for vertical wells when comparing
idealised logs in vertical and horizontal wells (White 1991)[Figure 3.5].

3.3.1 Tool errors

L WD and wireline comparison

There are two different types of tools used for formation evaluation in open-
holes: wireline and logging while drilling (LWD).
Open hole wireline logging is performed post-drilling and pre-casing, once mud
circulation has ceased. Wireline logs are accepted as the most reliable logging form of
continuous formation evaluation data but all wireline formation evaluation tools are
primarily designed for logging vertical wells. Nuclear wireline tools are all calibrated to
the same standard formations at the University of Houston notably all limestones, no
sandstone, dolomite or shale [Figure 3.6].

• Wireline logs from tools built by different companies can vary significantly as
conditions deviate from the standard conditions (Coope 1983).

LWD logging tools are an integral part of the bottom hole assembly (BHA)
which often consists of the bit, downhole motor, bent sub, stabilisers, drill collars and a
directional/inclination measuring tool [Figure 3.8]. The LWD tools primary purposes
are to act as drill collars and to remain in one piece while drilling, not as logging tools
(Allen et al. 1990; Woodhouse et al. 1991; Jackson et al. 1994; Prilliman et al 1995).

3-13
 Chapter 3: Horizontal well porosity anomalies

LWD tools have to cope with many more measurement perturbations (mud circulation,
chippings, vibrations) in comparison with wireline tools in the same well, although new
memory wireline tools can operate during mud circulation.

100- -

20 0 - -

300- -

Figure la - Deep resistivity response seen by sensor

Figure lb - Unique geological model derived from curve
logged across 300' of form ation.
In Fig. la assum ing n ear vertical incidence
Response changes at depths 'A ' and 'B '.
to bed boundaries.

Figures Ic to If - These models show four possible difTcrent Interpretations of Figure la assum ing wellbore is
horizontal. Here more data is needed to Identify the correct geological model.

———

F o rm a tio n
approxim ate
volume seen
by induction
tool.

approxim ate
volume seen
7ZZZZZZ TZZZZZl
F o rm a tio n
by density
tool.

Figure 3.5 Schematic to demonstrate the non-uniqueness problem when interpreting

horizontal well logs (Page 7 White 1991).

3-14
 Chapter 3: Horizontal well porosity anomalies

K 6 DIA,*! 3/8" STEEL DECK PLATE

* '4

FRESHWATER
CONCRETE

CORRUGATED PIPE

CARTHAGE LIMESTONE
(POROSITY s t.9% AVG-)

INDIANA LIMESTONE
0 9 % AVG. POROSITY)

AUSTIN LIMESTONE
(26% AVG. POROSITY)

CONCRETE

8" I D CASING

API Neutron Tool Calibration Pit

(Courtesy of Dresser Atlas)

Figure 3.6 Schematic of the nuclear tool calibration pit at the University of Houston
(Analysts 1997).

f~T R esistivity of th e z o n e
O R esistivity of th e w a te r in t h e z o n e
A W a ter s a tu ra tio n in t h e z o n e

A d jac en t b e d

Zone of
transition

lu d c a k e .

A d jac en t b e d

(In v asio n d ia m e te rs )

H ole
d ia m e te r

Figure 3.7 Schematic of the assumed logging environment (Gen3 Schlumberger 1999).

3-15
 Chapter 3: Horizontal well porosity anomalies

The LWD tool design varies greatly between manufacturers and there are no
common calibration standards, although Service Companies attempt to replicate
wireline responses. LWD tools are not calibrated to API standards directly because the
tools outside diameters are too large to run in the API calibration pits at the University
of Houston [Figure 3.6]. The calibrations are carried out in-house to a convenient
standard and then extrapolated to an apparent API value (Hutchinson et al. 1996).
Confusingly, LWD measurements are also called measurement while drilling
(MWD), but the important difference is that LWD measurements are of formation
evaluation quality (2-3 samples per ft) whilst MWD measurements are of lower
resolution (Theys et al. 1996). However, LWD and MWD are often used
interchangeably in the literature and to add to the confusion, measurement after drilling
(MAD) also uses the same tools as LWD and MWD. The author is only aware of MAD
measurements of formation evaluation quality, but it should not be assumed that MAD
measurements are always of formation evaluation quality.

Wireline drawbacks

The logging assumption that the formation is approximately isotropic

perpendicular to the borehole axis is more likely to be violated in horizontal wells
[Figure 3.7, Figure 3.27 and Figure 3.28] and conditions are more likely to deviate from
calibration conditions (primary calibration in vertical wells). For example, shoulder
beds can affect horizontal well values for hundreds of feet (Singer 1992).
When horizontal wells are logged with wireline tools the tools may not perform
optimally. Wireline tools also require a tool turner that orientates the tools so that
directional sensors run along the low side of the borehole pointing into the formation.
For vertical well logging the toolstring is pulled to the bottom of the borehole by
gravity. The process of logging in horizontal wells is time consuming and more likely
to damage tools and the borehole due to the need to push the tools down.

3-16
 Chapter 3: Horizontal well porosity anomalies

Modular
Connection

Power Section

Flex Joint —

Upper j
Stabilizer

24.1ft
Upper (7.4 m)
Transmitter —

Receiver
Antennas

G em m a R ay — G em m a
Inclinometer — Inclination A
Lower
Transmitter - if 18.4 ft
(5.6 m)

17.1ft
(5.2 m)

16.1 ft
Adjusts Ola ■ (4 3 m)
Kick Off S ub m ~J

Lower
Stabilizer

3.3 ft
(1.0 m)

i l i
Bit

Figure 3.8 Schematic of a BHA including an LWD tool (INTEQ 1999).

3-17
 Chapter 3: Horizontal well porosity anomalies

Wireline tools use skids and pads in combination with bowsprings and caliper
arms to eccentre the tool body to attempt good formation contact in both vertical and
horizontal wells. Pad/skid to formation contact or centralisation of a wireline tool also
may not be as reliable in horizontal wells as compared with vertical wells. Poor
pad/skid to formation contact normally produces increased apparent formation porosity
values. A non-zero density correction, DRHO, and out of gauge (different from the bit
size) caliper reading may also indicate standoff (poor pad/formation contact).
Logging speed affects not only the location of the measurements but also the
tool performance. Problems of the borehole bridging can cause the toolstring to become
stuck. Sticking is more problematic for wireline tools due to the greater formation
exposure time to mud. Normally sticking problems are automatically corrected for and
are not usually detrimental to the final wireline logs. Note: sticking in horizontal wells
is more likely because mud formation exposure is usually longer than with vertical
wells and the conveyance technique is physically larger than a wireline cable.

LWD drawbacks

LWD and wireline logs usually agree in vertical wells, mainly because LWD
tools remain centralised in the borehole while drilling (as during calibration). However
as Data Set 1 [See Chapter 4] demonstrated that MAD (after wireline logging in this
case) and wireline measurements can be quite different in vertical wells. LWD logs are
often the only formation evaluation data obtained from a horizontal well (Rider 1996).
The main advantage of LWD measurements is that mud invasion can be very
shallow and so the likelihood of measuring the virgin formation properties is greater
than that for wireline tools. LWD data can be of superior quality to wireline when
extensive formation damage and/or invasion occur while mud circulates for long
periods prior to wireline logging. LWD tools are often run as insurance logs or as a cost
saving measure in horizontal wells, since pipe conveyed wireline logging is expensive
because of the additional rig time as usually a dedicated well trip is required. However,
wireline memory logging negates a dedicated trip as the well is logged during well
conditioning.
For LWD tools the logging speed or ROP (rate of penetration) is dependent on
the drilling operation and the sample rate. The LWD data is recorded downhole in time
and converted to depth at surface. LWD data quality can be affected if the well is

3-18
 Chapter 3: Horizontal well porosity anomalies

drilled too fast as there will be too few samples per unit length (indicated by tick marks
on the log), then the values will not be representative of formation properties and a saw­
toothed log may result. Nuclear logs are statistical measurements and require a
minimum acquisition time per depth increment, sample rate -s- ROP, to produce a
reliable continuous log. Conversely, if the well is drilled too slowly, the tool memory
may be full before the zone of interest is logged completely or at all (Allen et al. 1990;
Brami et al. 1991; Hutchinson et al. 1996).
Poor quality LWD log data can result from poor time-depth data provided by the
drilling system. Often drilling depth systems are not robust or reliable; producing data
that contains errors such as communication errors leading to lost time-depth data;
driller’s depth corrections; and algorithm corrections to the depth. The physical position
of the depth encoder on the rig can also be a factor for accurate depth recording. The
problems with time-depth data result in misplacing of log data in depth which alters the
apparent bed thickness, also depth derived logs such as array resistivity and sonic logs
depend on correct depth indexing for their log processing.
LWD and wireline logs may not be the same because the tools’ spectral
responses are different and will diverge further when the measurement conditions and
formation become increasingly different from the standard conditions. If core data is
available core calibration of the logs should be used to provide a common point of
reference between LWD and wireline tool responses. This approach may suggest the
source of any differences between the logs. Another approach may be applied if tools
are run to log a specific horizon in a horizontal well. The tools can be calibrated within
the reservoir using a well-known marker bed (e.g. anhydrite bed) from knowledge
gained in previous logging runs, including core and log data in nearby pilot holes. If
there are differences between expected and actual values, and the borehole conditions
are good, then there may be a tool problem or new formation property information.
The two main constraints on LWD tools are that the sensors are behind an inch
of steel and direct formation contact is not possible. This imposes restrictions on LWD
porosity tool design, especially the density tool which is very sensitive to standoff.
Essentially LWD tools are mandrel tools (a smooth pipe) with varying degrees of
eccentring in the borehole (Allen et al. 1990; Day and Petler 1990). LWD tools are
normally centralised in vertical wells, but eccentred in horizontal wells because gravity
and the restrictions of the borehole prevent centralisation. To overcome some of these

3-19
 Chapter 3: Horizontal well porosity anomalies

problems LWD density tools use the minimum azimuthal count rate to provide the bulk
density measurement (Allen et al. 1990). This is based on the premise that errors due to
standoff tend to increase the count rate. The measurement is most probably made from
the low side of the borehole because the tool is weighted on to the low side of the
borehole, although instances of the LWD tools ‘riding’ the borehole to the right hand
wall during drilling (always clockwise) have been reported leading to the maximum
density values being recorded from the right hand quadrant.
Data recorded with azimuthal LWD neutron and density tools; after the Data Set
1 data presented in Chapter 4 was recorded but in the same field, demonstrated the
‘riding the borehole’ effect in horizontal wells. In fact the LWD density and neutron
tool ‘riding the borehole’ and the eccentred tool were the reasons for the poor LWD
porosity measurements detailed in Chapter 4. This demonstrates that new LWD logging
tool technology is being driven by the need to evaluate horizontal wells as well as by
the need to provide data as close to wireline quality as attainable. This is being
progressed, for example, by improvements in LWD tool design to cope with such
gravity related effects.

Summary o f LWD and wireline differences

The differences in recorded wireline and LWD values can be attributed to

(Jackson et al. 1994):

• Differences in sensor physics can produce different logs in identical conditions due
to the slightly different source-detector spacings, investigation depths and spectral
responses. In particular, neutron porosity can vary by 4-6pu due to epithermal,
thermal, capture gamma ray and lithology responses. LWD gamma ray tools
measure potassium rich values, so are more sensitive to shale and KC1 mud.
• The method of source-detector to formation contact will affect responses. Wireline
tools uses pad, skid or mandrel contact but LWD tools may use low absorption ports
through the stabilisers or can be run slick (without stabilisers). This is problematic
for the LWD density measurements particularly due to the effects of standoff.
• The basic reference standards are different.

3-20
 Chapter 3: Horizontal well porosity anomalies

• Environmental corrections must be applied to the wireline and LWD data to

common standard conditions.
• Filtrate invasion effects change between LWD, MAD and wireline logging runs.
• Formation alteration including clay alteration can have occurred.

LWD neutron porosity tool modelling

Day and Petler (1990) undertook numerical modelling of an LWD capture

gamma ray neutron porosity tool in a horizontal well (eccentred on the low side of the
hole) as compared with a vertical well (centred) to assess the likely error in apparent
porosity measured by the tool [Figure 3.9]. The modelled tool’s outside diameter was
7inches and the near and far detectors were assumed annular. The simulated formation
was 25pu fresh water saturated sandstone, the borehole diameter and borehole fluids
were altered and the centred and eccentred responses compared [Table 3.5].
One example of a LWD neutron detecting porosity tool was numerically
modelled [Figure 3.10] clearly demonstrated that eccentring a LWD neutron porosity
tool would decrease the measured porosity. This is as a result of the displaced fluid
from in front of the detectors as the tool is against the wall.

• LWD neutron porosity values in horizontal wells should measure lower values than
the same tool would in an equivalent vertical well (Day and Petler 1990).

Count rate calculated porosity (pu) for the eccentred tool calculated as if the tool were centred
Borehole Diameter (inches) Borehole Fluid Near Detector Porosity (pu) Far Detector Porosity (pu)
9 Water 22 24
10 Water 17 22
10 12 ppgMud 20 23
10* Water 15 20

* Calculated for a neutron detecting LWD neutron porosity tool.

Table 3.5 Porosity values for a modelled LWD capture gamma ray neutron porosity in
a horizontal well (Constructed from Figures 3 to 5 Day and Petler 1990).

3-21
 Chapter 3: Horizontal well porosity anomalies

Form ation Fo r m o t i o n

Borehole
Borehole
fluid

Tool

Tool

f l ui d

Calibration Conditions Horizontal Logging

Figure 3.9 Schematic of an LWD neutron porosity in a vertical (and calibration) well
and a horizontal well (Figure 2 Day and Petler 1990).

-
*
2.5 Near Detector

4)
(A <> d> O
c L ® O (?)
0 2.0
a : ------- Cr ------
Vt
<D : <d U)
o: 15 O o
-
+o
■>
u
■0 )
*->
1 .0
0) - Far Detector
Q
0.5
( ______ jJ,--- -O
-r r - 0' ~ i-i t”j : I _ *.

0 o ----- 1----- 1----- 1

----- 1----- 1----- 1----- 1----- 1----- 1---- 1----- 1---- 1— 1 I 1
180. 360. 540. 720.

Source Orientation

Figure 3.10 Detector response for an eccentred LWD neutron porosity (neutron
detecting) tool rotating in lOinch water filled borehole. The solid lines represent the
centred tool and the dotted lines represent the eccentred tool (Figure 5 Day and Petler
1990).

3-22
 Chapter 3: Horizontal well porosity anomalies

LWD and wireline comparisons

A comparison of LWD and wireline porosity tools in the Alaskan Endicott field
horizontal wells demonstrated that LWD tool stabilisation and sampling frequency are
important considerations for LWD logging (Cunningham et al. 1990), especially when a
long bit run is expected such as for MAD runs that are useful for monitoring invasion.
Table 3.7 compares wireline and LWD formation evaluations for the horizontal well 4
and Table 3.6 the formation properties used for the comparisons.

• The results showed LWD tools should be stabilised with full gauge stabilisers
immediately above and below the tool to reduce lateral drillstring motion and poor
centralisation, which would lead to erratic readings.

Positioning of the LWD tools as close to the bit as possible is also critical. The
closer the tools are to the bit, the shorter the formation exposure time to the mud and
thus invasion is shallower. However, the neutron tool (usually combined with the
density) is usually furthest from the bit in the LWD toolstring to allow source retrieval
if the LWD tools become stuck and to avoid formation activation that may affect
gamma ray readings. The neutron porosity comparison of wireline and LWD readings
demonstrated good agreement, except in shale and coal intervals [Figure 3.11] where
differences greater than 2-3pu were observed affecting the shale volume calculation.

• A different shale point would be required in the density and neutron porosity
calculation leading to potentially incorrect porosity and water saturation values.
V

Kekiktuk Permeability Gross Thickness Net/ d-Sw)

OOIP (MMstb) Porosity (pu)
Formation (mD) (feet) Gross (su)
3C 258 (23%) 18 300 180-250 0.33 82
3B 107 (9%) 18 150 90-170 0.21 80
3A 185(17%) 23 800 115-240 0.45 91
2B 256 (23%) 23 1500 30-160 0.90 90
2A 291 (26%) 22 1500 190-230 0.81 94
1 20 (2%) 15 150 105-280 0.19 78

Table 3.6 Properties of the formation used for the comparisons (Figure 2 Cunningham
etal. 1992).

3-23
 Chapter 3: Horizontal well porosity anomalies

Net Pay Average

Zone Average S« (su)
(feet) Porosity (pu)
3C Wireline 78.4 16.76 22.71
LWD 92.3 22.22 15.61

3B Wireline 37.9 19.71 15.61 •

LWD 40.1 15.49 23.64
3A Wireline 164.9 23.52 7.73
LWD 153.5 22.54 8.68
2B Wireline 97.5 24.26 24.28
LWD 96.7 24.60 36.02
2A Wireline 163.2 21.41 16.28
LWD 165.2 22.43 22.80

OIL Wireline 212.8 23.73 7.92

LWD 195.3 22.87 8.28
GAS Wireline 141.3 17.14 23.73
LWD 142.8 22.74 20.31

Table 3.7 LWD/Wireline porosity comparison in well 4 [non-stabilised density]

(Table 1 Cunningham et al. 1992).

The LWD density tool measurements maybe compared with wireline

measurements in Table 3.7 and Figure 3.11. In zones 3C and 3B, there majbe
differences of up to 5pu between wireline and LWD readings potentially affecting shale
volume and hence porosity values. Often the large differences correspond to non-pay
shaley intervals with up to 10 inches of borehole enlargement. However, horizontal well
12 demonstrated that the LWD density was reading too high bulk density (low porosity)
when checks were made against wireline and core measurements.

• Clay hydration between the LWD density and wireline measurements could not be
ruled out and could have caused the departure of the logs (Cunningham et al. 1990).
• Non-parallel standoff from LWD and/or wireline tool running across a chord could
also result in the LWD and wireline differences (Samworth 2000).

A dual source LWD density was run in another well in tandem with a single
source tool and a LWD neutron tool, although no wireline tools were run. The logs
demonstrate that the dual source density readings were less erratic and in close
agreement with the neutron porosity values in shale free formations [Figure 3.12] due to
improved count rate statistics and standoff correction available from dual source tools.

3-24
 Chapter 3: Horizontal well porosity anomalies

Cunningham et al. (1992)’s recommendations included well-stabilised LWD

density (preferably dual source) tools with pseudo-caliper corrections can provide
reasonable results, except in boreholes in which rugosity and washout are likely to be
an issue. However, the quality of LWD tool measurements is as yet far from resolved
while significant differences between LWD, wireline and core measurements persist.

• The author’s experience of LWD logs are that they are poor quality when compared
with wireline data, the main differences being depth of investigation, bed boundary
resolution and noise rejection of the LWD data.

LWD tools from different manufacturers (Teleco, Sperry Sun, Schlumberger)

were compared with wireline tools from different manufacturers (Halliburton,
Schlumberger) in vertical wells of the Alaskan Kuparuk river field (Sakurai et al. 1992).
The Kuparuk river formation units A and C are the main producing units within the
formation. Wireline and LWD log response comparisons were used to assess the quality
of the LWD measurements for formation evaluation purposes.

• Differences in tool design and in logging environments between wireline and LWD
lead to differences between the recorded measurement values (Sakurai et al. 1992).
• Primary LWD tool characterisation to API standards and outside of API standard
conditions could be poor when compared with wireline tools.

Gamma ray measurements were used to calculate clay content in the A sands
and as an input for mineral determinations with other logs for the C sands. Figure 3.13
demonstrates that the gamma ray readings from LWD Company 1 were consistently
higher than the wireline measurements, possibly due to incorrect calibration procedures
(Sakurai et al. 1992).
Wireline and Company 1 LWD neutron porosity comparisons from vertical well
4 found reasonable agreement [Figure 3.14]. However, wireline and Company 3 LWD
neutron porosity comparisons in vertical well 5 show that large differences are apparent
in the C sands in vertical well 5 where heavy minerals are present [Figure 3.14]. The
heavy minerals increase the formation density, which decreases the count rates of
capture gamma rays used by Company 3 to calculate neutron porosity.

3-25
 Chapter 3: Horizontal well porosity anomalies

VyvJ

O
J-
m oX
<
C3
=I1
“•
c'Q

I
f V ^ T |(j i s i / w ' V

ul
ill
Ul2

3
m
_
<U
<b
o
<b
o
c
c3
d
C>
O
O
O
O
<U *— CNJ C m id
u_ x x :< >C X X

“1 <=
om
\il 1
HL LI-
Flf
CQ
Hn
j l l

<
1 ft) )

CO m

Figure 3.11 Wireline and LWD neutron and density (non-stabilised density)
comparisons for horizontal well 4. Wireline and LWD density (stabilised density)
comparison with core porosity for horizontal well 12 (Figures 12 and 13 Cunningham et
al. 1992).

3-26
 Chapter 3: Horizontal well porosity anomalies

RHOB 1 g m /c c
THOB 2~g~m7c~c~

R H O B (sin g le )

R H O B (D ual)

■x300-

x400-

Figure 3.12 Tandem LWD tools run of single and dual source non-stabilised density
tools, neutron was also run (Figure 14 Cunningham et al. 1992).

150

120
Q.
<
c5ok
DC

E
00
e> ■a ■
=5

0 30 60 90 120 150
Company 2 - Gamma Ray ; API

Figure 3.13 Crossplot of wireline and MWD gamma ray by Company 2 from vertical
well 5 (Figure 2 Sakurai et al 1992).

3-27
 Chapter 3: Horizontal well porosity anomalies

,6 0 __________________________Q
CN0 P :U .
iL£5__________________ 2JL5
RHOB 1 g m /c c
jl s q ilJL5__________________ 2JL5
GR API RHOB 2 g m /c c

R H O B (sin g le)

RH O B(D ual)

-x 2 0 0 -

■x300-

•x 4 o o-

Figure 3.12 Tandem LWD tools run of single and dual source non-stabilised density
tools, neutron was also run (Figure 14 Cunningham et al. 1992).
150

120
»
Q.
< . •
CO
oc llp V

! 60 4 0
O
*

30

/ 30 60 90 120
Company 2 - Gamma Ray ; API
150

Figure 3.13 Crossplot of wireline and MWD gamma ray by Company 2 from vertical
well 5 (Figure 2 Sakurai et al 1992).

3-27
 Chapter 3: Horizontal well porosity anomalies

Histograms of density porosity difference (wireline-LWD) separated by

company and sand unit show that Company 1’s measurements were in closer agreement
with wireline than Companies 2 and 3 [Figure 3.15](Sakurai et al 1992). LWD density
porosity measurements in the C sands showed larger scatter than in the A or B sands for
all companies [Table 3.9]. The scatter was generally greater for LWD density
measurements than wireline measurements.

• LWD tools tended to over correct for borehole conditions due to the strong
dependency of the short spaced detector response (Sakurai et al 1992).

The differences between wireline and LWD porosity measurements are given in
Table 3.8 and Table 3.9 for the density and neutron tools. The investigation discovered
that the LWD tools of Company 1 were in closest agreement with wireline tool
responses. A mismatch of up to 8pu could be expected from some of the tools (Sakurai
et al 1992). When compared with their wireline counterparts LWD density values were
twice as poor as LWD neutron porosity values, which broadly agrees with the results of
numerical modelling studies (Day and Petler 1990).

• LWD and wireline differences demonstrated that the porosity anomalies were not
solely tool based (Sakurai et al 1992) but also due to environmental effects, i.e.
borehole conditions and shale alternation due to hydration.

LWD Company 1 2 3 LWD Company 1 2 3

Well 4 5 5 Well 4 5 5
C Sand C Sand
Mean Difference (wireline - -0.3 -0.6 -2.9 Mean Difference (wireline - -0.7 4.2 -3.3
LWD) Neutron Porosity (pu) LWD) Density Porosity (pu)
Standard deviation (pu) 1.8 1.6 3.6 Standard deviation (pu) 3.2 8.1 6.1
A/B Sands A/B Sands
Mean Difference (wireline - -0.3 0.7 0.2 Mean Difference (wireline - -0.5 -0.9 -0.9
LWD) Neutron Porosity (pu) LWD) Density Porosity (pu)
Standard deviation (pu) 1.4 1.6 1.6 Standard deviation (pu) 1.2 2.8 2.9

Table 3.8 Wireline/LWD neutron Table 3.9 Wireline/LWD density tool

porosity tool comparison (Table 3 Sakurai comparison (Table 4 Sakurai et al 1992).
etal 1992).

3-28
 Chapter 3: Horizontal well porosity anomalies

..4 0

2 30 _i 30

C sand

• A/B sa n d s

10 20 30 40 50 60
10 20 30 40 50 60
C om pany 3 - NPHI : P.U.
Com pany 1 - NP sa n d ; P.U.

Figure 3.14 Crossplots of wireline and MWD neutron porosity by Company 1 from
vertical well 4 and by Company 3 in vertical well 5 respectively (Figures 7 and 8
Sakurai et al 1992). Reasonable agreement is found for vertical well 4, but large
differences are apparent in the C sands in vertical well 5 where heavy minerals are
present. Company 3 uses capture gamma-ray methods.

Company 1 C om pany 2 Company 3

C sand

A/B Sands

Figure 3.15 Crossplots of wireline neutron porosity against MWD neutron porosity by
Company from vertical well 5 (Figure 6 Sakurai et al 1992). Data to the left of centre
represents an overestimate of porosity by the MWD neutron porosity tool.

3-29
 Chapter 3: Horizontal well porosity anomalies

Standoff

Standoff corrections for LWD density and neutron tools have always been
problematic (Moake et al. 1996). At the introduction of LWD density and neutron tools
it was generally thought that the effects of standoff would be negligible, based on the
assumption of no significant borehole enlargement prior to LWD logging. This
assumption was soon known to be incorrect and that standoff was particularly
problematic in horizontal wells. Wireline density and neutron tools are much less
susceptible to standoff effects than their LWD counterparts due to the use of a caliper
and/or bowspring, except possibly in severely eroded/elliptical holes when wireline
tools may not sit along the major axis.
Four techniques have been developed to reduce the standoff errors (Best et al.
1990; Holenka et al. 1995; Spross et al 1995; Moake et al. 1996). Best et al. (1990) used
a histogram of count rates over a time interval that included several revolutions of the
tool. Standard deviations of the near detector count rates were compared with expected
count rates at various values of standoff. This provided an indirect estimate of the
standoff, an azimuthal average density value and a maximum density value. This
method was an improvement to no correction at all [Figure 3.17], but even with only
modest standoff low bulk density values resulted [Figure 3.16]. Spross et al (1995) used
a histogram of count rates over a time interval that was much shorter than one
revolution of the tool. Again the standard deviations of the count rates were compared
with expected count rates at various values of standoff, but the near and far count rates
were summed for this purpose. If the ratio of the two standard deviations was sufficient
the count rates are binned, dependent on the ratio. A spine and ribs analysis was then
used to evaluate which bin represents the least amount of standoff, which was then used
as the value of RHOB. The Spross et al (1995) method was still indirect but improved
the standoff correction at the expense of throwing away much of the data [Figure 3.17].
Holenka et al. (1995) used four bins based on tool orientation in the borehole. It was
assumed that the smallest standoff occurred in the ‘down’ direction, not always the
case, and the appropriate DRHO indicated the likely error for each quadrant. Generally
the method was an improvement on the two previous methods [Figure 3.17]. Moake et
al. (1996) uses ultrasonic transducer measurements of standoff, which were then used to
weight the count rates, exponentially decreasing the weight with increasing standoff.
The main improvement in this method was that the standoff was measured directly, so

3-30
 Chapter 3: Horizontal well porosity anomalies

one would expect values that are more accurate. However, any standoffs greater than
linch is more than any wireline or LWD correction algorithm can correct for.
Moake et al. (1996) added two new correction values, Corr P and Corr M, which
summed together equal the familiar density correction log. Corr P measures the density
contrast and is usually positive, while Corr M measures the PEF contrast and is usually
negative. These values are useful in heavy muds {high PEF values [Figure
3.18](Minette 1996)} in which DRHO can be zero but Corr P and Corr M would be
large and have opposite sign. The difficulty is in how to compensate for these high PEF
values because of similarity of the slopes of the formation and heavy mud PEF as
defined by the hard count rate versus the hard/soft ratio [Figure 3.19].
The Evans et al. (1999) neutron compensation method, for borehole invariant
porosity, effectively makes the neutron porosity measurement insensitive to the first
1finches of the tool’s depth of investigation. This is achieved by modification of the
far detector count rate by the use of a functional relationship with the near detector. The
ratio of the modified far count rate and the near count rate is then used to calculate
porosity.
2500

2000
1000
Near _ Average
Detector - of one
1500 Far - oscillation Standoff
Count Count (inches)
Rate Rate 1.75
1000

Far 2.60 g/cc

soo Detector

100 JL
0 0.5 1 1.5 2 1000 3000
Standoff (inches) Near Count Rate

Figure 3.16 Illustration to show the non-linear effect of rotation on the LWD density
measurement in a 2.6g/cm3 formation with 131bm/gal mud in a lOinch well (a) count
rate versus standoff and (b) spine and ribs format (Figure 1 Moake et al. 1996). The
double arrow in (b) indicates the oscillations along the rib when the LWD 6.75inch tool
is lying on the bottom of a horizontal well. The solid circle represents the average count
rates measured during one oscillation.

3-31
 Chapter 3: Horizontal well porosity anomalies

0.4 0.4 0.4

(a) First (b) Second (c) Quadrant
0.3 Statistical Statistical Technique
Technique / 0.3 Technique / 0.3 /
0.21- /
Density
Error
' #/ fjs Density 0 2
Error
Density nP
Error
J /
0.1 i /
(g/cc) /! (g/cc) to /
(g/cc)

0.0
0.1 / I
J 0.1 /
5 Quadrant
CO
- 0.1 ■— i—— «— 0.0 S - J . 0.0
8 9 10 11 12 8 9 10 11 12 8 9 10 11 12
Hole Diameter (inches) Hole Diameter (inches) Hole Diameter (inches)

Figure 3.17 Density errors associated with various processing techniques (a) first
statistical method (Best et al. 1990), (b) second statistical method (Spross et al 1995)
and (c) quadrant method (Holenka et al. 1995) from (Figure 6 Moake et al. 1996). Note:
Errors from average data with no special processing (raw) are compared with the errors
presented. The environmental conditions are as Figure 3.16.

Mudcake Response
100 . » ■IT T ’T-T T—T— T— —■■■nr—•«—
■— —Heavy Mud Hard
' — — Heavy Mud Hard/Soft
y*-
c
. ——— Light Mud Hard *
0) .„......— Light Mud Hard/Soft /*
E
£3
/ '
<0 r
<0 10
4> »
» /
5 »
a /
*

■a
\

5
\

-
E /
#

o /

' * 1.0 ? .. .

• . -
0.5
.1 J2 .3 .4 .5 .6 ,7 .8 .9 1.0
Mudcake Thickness (in.)

Figure 3.18 Plot to demonstrate the effects of light (/?=1.3g/cm3, PEF=0.3B/e) and
heavy (/7=2.5g/cm3, PisF=101B/e) mudcake on the spectra obtained from a 2.7g/cm3
(P£F=2.57B/e) aluminium formation (Figure 7 Minette 1996).

3-32
 Chapter 3: Horizontal well porosity anomalies

Hard v s . H ard/Soft Ratio

1.20 c
. ' Formation Po
1.15 . . . . . . . . Heavy Mud
—— — Light Mud
1.10

X 1.00

Z .85

.80

.75

.70
1.0 10 100
H ard/Soft Ratio

Figure 3.19 Plot to demonstrate that the density measurement is less affected (similar
slope to the formation response) than the PEF measurement for ‘heavy’ muds (i.e. mud
with high PEF) (Figure 8 Minette 1996). The effects for light mudcake response on the
spectra are due to both the density and PEF of the mud, a different slope to that of the
formation or heavy mud. The heavy, light and formation values are as for Figure 3.18.

Gamma ray tools

Thorium/Potassium (Th/K) crossplots are often used for mineral identification

when natural gamma ray spectrometry (NGS) logs are available [Figure 3.20](Hurst
1990). At present no LWD NGS tools exist and, considering the effect on the spectral
response of the tool by the drill collars (makes logs K rich), no LWD NGS tools are
likely in the near future (Sakurai et al 1992). However, the use of such crossplots is
limited, since there is no theoretical basis for their use for mineral identification [Figure
3.21](Hurst 1990). Figure 3.21 demonstrates that the mineral boundaries of Figure 3.20
typical of the crossplots found in Service Company chartbooks are fictional at best.
Mineral identification is made significantly more difficult in sandstones with moderate
K-feldspar contents, as is common in the North Sea, for example, the associated error of
the proportion of K-feldspar is such that the error would completely overwhelm any
measurement of illite content.

• The measurement of illite could be potentially useful for evaluating reservoir

quality because of the effect of illite on permeability (Hurst 1990).

3-33
 Chapter 3: Horizontal well porosity anomalies

Possible 100% kaolinite,

montmorillonite,
illite “day line"

Kaolinite

r-^40%
/ mica

•30% glauconite
Glauconite
Feldspar
Potassium evaporites, ~30% feldspar
0 1 2 3 4 5

Potassium {%)

Figure 3.20 Thorium versus Potassium plot for mineral identification (CP-19
Schlumberger 1999).

Th versus K
— Th:K 0.19 — Th:K 1.0 Th:K 1.16 — Th:K 1.2
— Th:K 3.2 — Th:K 7.1 — Th:K 10 Th:K 470

K a olin ite

AQ- M u sc^
0 1 2 K (%) 3 4 5

Figure 3.21 Thorium versus Potassium plot (Redrawn Figure 6 Hurst 1990). Note: The
considerable areas of overlap between potassic minerals and the broad kaolinite field.
Kaolinite normally has negligible potassium content.

3-34
 Chapter 3: Horizontal well porosity anomalies

Filtering

Filtering of log data effects not only character of log curves, but also the
interpretation of the curves themselves. All log curves are filtered for several reasons,
such as resolution matching of different detector spacings and noise removal, but also
most logs have presentation filters applied to them to improve the visual appearance of
the curves. Typically a 1’ and Wz convolved moving average filter is applied to the
porosity logs. The filter retains the fine bed resolution data (to 1’) whilst removing most
of the high frequency random noise in the data. If faithful comparisons between LWD
and wireline data are to be made presentation filters need to be removed. Unfortunately
obtaining data without presentation filters applied is extremely difficult, as these filters
are often not optional in the log processing of the Service Companies. When comparing
wireline and LWD logs filtering the data to the same wavelength can be useful so that
the logs can be compared at the same resolution, usually 2ft and can help with depth
correlation because the filter matches the vertical resolutions (Schroeder et al. 1991).

Summary o f tool errors

Tool errors for porosity tools largely result from differences in detailed tool
design and toolstring configuration whilst logging. Calibration and modelling of LWD
tools is often insufficient to replicate wireline responses and common standards for all
logging tools need to be established. Poor understanding and control of logging tool
position within the borehole whilst logging leads to poor porosity logs particularly in
horizontal wells and wells in which lithologies which deviate significantly from
calibration points are encountered.

• Wireline logs from tools built by different companies can vary significantly as
measurement conditions deviate from the standard conditions (Coope 1983; Sakurai
et al 1992).
• Differences in sensor physics can produce different wireline and LWD logs in
identical conditions due to the slightly different source-detector spacings, depths of
investigation and spectral responses (Jackson et al. 1994; Sakurai et al. 1992).

3-35
 Chapter 3: Horizontal well porosity anomalies

• Method of source-detector to formation contact will affect responses. Wireline tools

use pad, skid or mandrel contact but LWD tools may use low absorption ports
through the stabilisers or can be run slick. LWD density measurements are
particularly problematic due to the effects of standoff (Jackson et al. 1994), over
correction for standoff is due to the strong dependency of the short spaced detector
response (Sakurai et al 1992).
• Basic reference standards are different (Jackson et al. 1994).
• Environmental corrections must be applied to the wireline and LWD data to
common standard conditions (Jackson et al. 1994).
• LWD neutron porosity values in horizontal wells should measure lower values than
the same tool would in an equivalent vertical well (Day and Petler 1990).
• The results showed LWD tools should be stabilised with full gauge stabilisers
immediately above and below the tool to reduce lateral drillstring motion and poor
centralisation, which would lead to erratic readings (Cunningham et al. 1990).

The author’s experience of LWD logs are they are often poor quality when
compared with wireline data, the main differences being depth of investigation, bed
boundary resolution and noise rejection of the LWD data. Other problems occur such
as:

• A different shale point would be required in density and neutron porosity

calculation where tool responses differ. Failure to recalculate the shale point may
lead to potentially incorrect porosity and water saturation values.
• Non-parallel standoff from LWD and/or wireline tool running across a chord could
result in the LWD and wireline differences (Samworth 2000).

Jackson et al. (1994) provides a good discussion of the differences of LWD and
wireline while Brami et al. (1991) has a detailed description of the differences in LWD
tools offered by several Service Companies.

3-36
 Chapter 3: Horizontal well porosity anomalies

3.3.2 Borehole conditions

Borehole conditions that affect logging tools include mud density changes, mud
additives (Barite, KC1 and Hematite), rugosity, hole size, temperature, pressure,
borehole salinity and invasion (Hutchinson et al. 1996). Before comparing different
logs they must be corrected for the same borehole conditions (Jackson et al. 1994).

Hole size

LWD density tools are critically dependent on good standoff corrections and the
density contrast between the formation and mud (Spross et al 1995). For LWD density
tools, if standoff is small and the tool size to hole volume ratio is not significantly
increased by hole enlargement (mud volume), the standoff effect is small (Minette et al
1995). Several LWD tool sizes are required for the appropriate hole size compared with
wireline tools where only one tool size is used for all hole sizes, although different
Service Companies have different sized wireline tools from one another.

• The standoff effect on LWD density tools may be especially severe when the
drillstring is sliding; the standoff algorithm may become ineffective because of the
dependence on the short spaced detector values from all azimuths (Allen et al. 1990;
Bestetal. 1990).

LWD neutron porosity tools are designed to run centralised and are more
affected by the hole condition (particularly standoff) than the eccentred wireline
neutron porosity tools (Allen et al. 1990). The formation is effectively removed from all
around the LWD tool but is only on one side of the wireline tool and leads to the LWD
tools reading too high apparent porosity (Brami et al. 1991). An LWD neutron porosity
tool is affected by approximately 5pu per inch standoff (Jackson et al. 1994).

• For a 40pu formation and a 2inch standoff, the error for a wireline tool is
approximately 1.5 to 3.0pu, but approximately 12pu for a LWD tool (Allen et al.
1990).

3-37
 Chapter 3: Horizontal well porosity anomalies

Deviated holes are often oval and can cause the LWD density tools to make
inconsistent formation contact, especially when the drill string is sliding and the sensors
can be facing into the mud column (Brami et al. 1991). No or slow rotation of LWD
tools may lead to standoff algorithms being unable to correct for oval holes (Day and
Petler 1990). A dual source LWD density tools improves density calculation due to
better count rate statistics and improved compensation (Cunningham et al. 1992).

• Wireline tools can also suffer from poor formation contact if the toolstring rolls
over and the pad is orientated into the mud column. This results in under-gauge
caliper values, because the tool weight partially collapses the caliper and possibly
non-zero DRHO values (Samworth 2000).

Typically LWD density-neutron crossover is greater in gas than for comparable

wireline tools, because invasion is usually less extensive (Carpenter et al. 1997). Time
after bit is a useful indicator for possible washouts or if the log was recorded during a
pipe or conditioning trip. Often time after bit and LWD top quadrant (ADN tool)
DRHO correlate because barite mud adversely affects top density values, and often only
the bottom density value is reliable. Thus an understanding of the effect of standoff on
PEF in barite muds [Figure 3.22] can assist interpretation, as can LWD density and
PEF images [Figure 3.26].

• The selection of logging tools has a dramatic effect on the data quality recorded.
• The selection of logging tools should be based on the expected borehole conditions
[Figure 3.23](Allen et al. 1990).

Formation compaction

Formation compaction, caused by the drill pipe, may occur immediately below a
keyseated borehole (low side borehole enlargement) providing a place for debris to
collect. A positive DRHO would be measured (Cuddy et al. 1994)[Figure 3.24] and
would thus reduce log quality (Nieto et al. 1995).

3-38
 Chapter 3: Horizontal well porosity anomalies

12.0 ppg

9.5 ppg

0.4 0.5 0.6 0.7 0.8 0.9

Standoff (in.)

Figure 3.22 Plot of PEF versus standoff for different mud weights (Figure 1 Carpenter
et al. 1997).
CDN Hole
.
Wireline \
Condition
Hole Condition
oo0 0 O
Smooth and
O

in-gauge

with mudcake
O
E xcellent L_ -
aareem enti
Smooth and
enlarged

Smooth,
0 Rotationall

enlarged
with mudcake 0 j re q u ire d ] I

Smooth
and altered Time lap se density
formation
density
..............
© interpretation p ossible

Enlarged
and rugose o .

Enlarged, Sonic p°roslt>

rugose prefered
and altered
formation
density

Figure 3.23 Guide for selecting logging tool type depending on expected borehole
conditions (Allen et al. 1990). Note: the CDN is Schlumberger’s LWD Combinable
Density Neutron tool.

3-39
 Chapter 3: Horizontal well porosity anomalies

Pipe Groove and Cuttings

possible compaction

Groove and Cuttings Tool Orientation

Figure 3.24 Possible horizontal well I conditions redrawn (Figure 16 and 17
Cuddy et al. 1994).

3-40
 Chapter 3: Horizontal well porosity anomalies

Rugosity

Rugosity introduces variable standoff and pad tilt for wireline tools that is not
corrected for by standard corrections (Jackson et al. 1994) and may be due to spiral
grooves caused by drilling (Norve et al. 1989; Betts et al. 1990). Spiralled boreholes
were encountered in logs from Data Set 1 [See Chapter 4]. In rotation the LWD density
tool, if the average compensation is >0.1g/cm3 will under compensate (Minette et al.
1995). In sliding, if the LWD density tool is measuring the horizontal well floor, a large
DRHO may result. There is no preferred LWD density tool orientation in vertical wells.
The BHA stabilisation assumes the LWD tools are rotating about the borehole
centre and parallel to the borehole wall (Cunningham et al. 1992). Full gauge stabilisers
above and below LWD tools help to centralise the sensors and dampen the lateral BHA
movement (Cunningham et al. 1990), however BHA modelling shows lateral BHA
motion is possible (Delafor 1984). Full gauge stabilisers with sensors behind them may
not totally solve these problems and results in highly forwarded focused density
measurements. In addition, drillers are usually unhappy drilling with stabilisers while
building (increasing hole inclination) because of the increased potential for BHA
sticking, borehole and drill string damage.

• BHA motion is not strictly rugosity but the motion will manifest itself as variable
standoff as rugosity does (Betts et al. 1990).
• Debris in horizontal wells collects along the borehole floor and can introduce
variable standoff (Woodhouse et al. 1991).

Betts et al. (1990) demonstrated that stabilisers caused LWD density (and
neutron porosity) log to vary by ±0.2g/cm3 (±20% porosity) on a 3ft wavelength with
rugosity no more than 0.2inches (6mm). Conventional filtering removed too much low
frequency response making the log too smooth (Betts et al. 1990). The BHA was
possibly wobbling between the first and second stabilisers or the first stabiliser and bit.
The period was the same in the LWD density and caliper curves and unaffected by
change in rate of penetration. A Weiner filter and 3ft-notch filter was used to remove
periodicity. Another period of 75 ft was observed, but no plausible explanation was
found.

3-41
 Chapter 3: Horizontal well porosity anomalies

ADN RH06 DYNA |C70063)

HarizontalScate 1Q. Q 0 45 -(1151
________________ [ fT3/ft3)________________ j
Coentaoon Top of Hofe
RC'BU
histogram equaized
1 95 296
( q/cm 3l _______________
RHOB Image .............................. FOBL............................... i
195...............................................................*295i
(q/cm3) ;
MD ___________ ROBR _______________
1 : 200 1 95 ' 295
ft (g/an3) ___________________( g /c m 3 )____________________
image onentanon ROBB
R B L

1L050

12000

12050

Figure 3.25 LWD density image of a spiralled borehole above 12012ft, below a near bit
stabiliser was added to prevent spiralling (Figure 3 Maeso et al. 1999).

3-42
 Chapter 3: Horizontal well porosity anomalies

D E N SIT Y

PHOB RHQ8
(G/C3) lG vC 3l

Figure 3.26 Density and PEF images in a spiralled borehole with clockwise sense
above xl70ft, but anticlockwise sense below (Figure 5 Carpenter et al. 1997).
 Chapter 3: Horizontal well porosity anomalies

The bent sub used to alter hole angle enlarged the hole when the BHA was
rotating (Betts et al. 1990; Nieto et al. 1995). The offset that resulted in the low bulk
density values was removed by taking half the near/far count rate difference to give a
corrected count rate. The bulk density and photoelectric factor values were recalculated
and the neutron porosity log was reprocessed in a similar manner (Betts et al. 1990).
The first stabiliser is normally V/tr^ft behind the bit and acts as a node (Betts et al.
1990). The groove wavelength produced was comparable with the source to detector/s
spacing/s and so interfere with density and neutron measurement.

• The periodicity is twice the bit to first stabiliser distance and the modulation
depends on the weight on bit and bent sub angle (Betts et al. 1990).

Borehole spiralling by a drill bit can produce periodic variation of LWD density
logs/images (Maeso et al. 1999), usually at a wavelength of approximately 3ft, caused
by oscillation of the bit and stabiliser [Figure 3.25]. However, the mechanism is not
fully understood (Carpenter et al. 1997), because observed instantaneous sense reversals
of the variations cannot be explained [Figure 3.26]. Inclusion of a near bit stabiliser can
prevent the oscillations. Modelling demonstrated that LWD density images allow
carbonate nodules to be differentiated from carbonate beds, which may avoid expensive
wells being drilled to penetrate the apparent ‘carbonate bed’ (Maeso et al. 1999).

Mud type

Norve et al. (1989) studied the effects of mud type with LWD and wireline
logging in vertical wells from a Norwegian North Sea field and demonstrated that LWD
measurements are sensitive to mud type. The field used was comprised of low
resistivity pay zones, in which accurate porosity determination is critical for a reliable
Swcalculation [Table 3.10].

• The LWD porosity tools performed best in oil based mud (OBM), followed by NaCl
mud and were poorer still in KC1 mud (Norve et al. 1989).

3-44
 Chapter 3: Horizontal well porosity anomalies

This was possibly due to the mud Cl content and the differences in spectral
responses of the tools. The best LWD logs were obtained when the rate of penetration
was less than 40m/hr (120ft/hr) (Norve et al. 1989). Mud additives can affect the quality
of log derived porosity values. The effect of barite and hematite is negligible on PEF
for mud weights less than 10.51bm/gal (1.26g/cm3)(Jackson et al. 1994). There is
usually less barite or hematite in the mud when LWD tools are run than wireline tools
because of the settling of the heavy minerals over time.

• The author is aware of numerous wells with mud weights well below 10.51bm/gal
(1.26g/cm3) especially OBM in which barite can have substantial effects on PEF
logs to the point that they are rendered useless.

Pay Type Net/Gross (Frac.) Porosity (pu) Sw (su)

LWD Wireline LWD Wireline LWD Wireline
NaCl mud
Oil 0.35 0.59 21.7 24.7 46 38
Water 0.48 0.33 17.4 19.0 81 91
KC1 mud
Gas 0.90 0.94 25.3 20.8 20 13
Oil 0.56 0.54 23.6 20.6 39 36
Water 0.80 0.71 22.8 20.5 89 84
Oil based mud
Oil 0.56 0.56 22.3 22.3 10.5 10.5

Table 3.10 LWD/Wireline comparison against mud type (Tables 2 to 4 Norve et al.
1989).

3-45
 Chapter 3: Horizontal well porosity anomalies

Summary o f borehole conditions

Borehole conditions affect porosity logs. Poor conditions can result in poor logs,
especially from LWD porosity tools and affect the control of logging tool position
within the borehole. Horizontal well conditions are likely to be worse than a
comparable vertical well due to increased stress on the borehole from overburden
pressure and debris collecting along the length of the well.

• In sliding mode, the standoff effect on LWD density tools may be especially severe;
the standoff algorithm may become ineffective because of the dependence on the
short spaced detector values from all azimuths (Allen et al. 1990; Best et al. 1990).
• Debris in horizontal wells collects along the borehole floor and can introduce
variable standoff (Woodhouse et al. 1991).
• For a 40pu formation and a 2inch standoff, the error for a wireline neutron tool is
approximately 1.5 to 3.0pu, but approximately 12pu for a LWD tool (Allen et al.
1990).
• The selection of logging tools should be based on the expected borehole conditions,
(Allen et al. 1990), though a chart should to devised based on regional best practice.
• The spiral rugosity periodicity is twice the bit to first stabiliser distance and the
modulation depends on the weight on bit and bent sub angle (Betts et al. 1990).
• The LWD porosity tools performed best in oil based mud (OBM), followed by NaCl
mud and were poorer still in KC1 mud (Norve et al. 1989).

The author’s experience of LWD logs indicates they are often poor quality when
compared with wireline data, the main differences being depth of investigation, bed
boundary resolution and noise rejection of the LWD data. Most of the degradation in
borehole quality occurs prior to LWD logging and in many cases there is little
difference in borehole conditions. The wireline porosity tools normally will cope better
with the poor borehole condition providing better logs. However in some areas this is
not so due to regional stresses (e.g. The Caspian), where washout and the risk of
borehole collapse are such that wireline tools are seldom used. Other problems occur
such as:

3-46
 Chapter 3: Horizontal well porosity anomalies

• Numerous wells with mud weights well below 10.51bm/gal (1.26g/cm3) especially
OBM in which barite can have substantial effects on PEF logs to the point that they
are rendered useless.
• Wireline tools suffer from poor formation contact if the toolstring rolls over
resulting in poor density values and under-gauge caliper values (due to the tool
weight partially collapsing the caliper) and possibly non-zero DRHO values
(Samworth 2000).
• BHA motion will manifest itself as variable standoff similar to rugosity.

Jackson et al. (1994) provides a discussion of the effects of differences of LWD

and wireline values on formation evaluation and Allen et al. (1990) provides a guide for
the porosity logging method to be employed dependent on the borehole conditions
expected.

3-47
 Chapter 3: Horizontal well porosity anomalies

3.3.3 Altered formation conditions

Formation evaluation attempts to calculate the virgin formation conditions from

the tool responses. However, the very process of drilling a borehole alters the formation
conditions. The effect of the altered conditions is to change the log responses
accordingly. The main processes that affect the formation condition are mud invasion,
fractures, formation alteration and bedding.

Invasion

Mud filtrate invasion alters the formation fluid content displacing connate water,
oil and gas to irreducible saturation when invasion is complete. If the tool measured
volume and the invaded volume coincide the measured formation properties will not be
representative of the virgin conditions, unless the mud filtrate is accounted for. In
vertical wells mud invasion is assumed piston-like with the invasion front moving out
radially from the borehole.

• Gravity and permeability anisotropy have a combined effect on mud invasion in

horizontal wells that may produce non-uniform invasion profiles, consequently
affecting log measurements (Woodhouse et al. 1991).

The primary control on invasion is formation permeability and secondly

mudcake permeability (Harris et al. 1993). Note that porosity is also a major factor as
it’s a measure of the fluid volume capacity of the rock. Invasion is insignificant due to
effective mudcake usually in 8-10hrs after drilling with a typical mudcake permeability
of 0.00006mD to 0.00013mD (Woodhouse et al. 1991). Deep invasion can be caused by
excessive mud (pressure) overbalance and placing the LWD tools too far from the bit
results in excessive formation exposure time (Brami et al. 1991).

• Generally the invasion process is static by the time wireline tools are logged (>8-
lOhrs after bit penetration, equilibrium having been established).

3-48
 Chapter 3: Horizontal well porosity anomalies

An important factor is that the logging suite can measure the saturation of the
various formation fluids and account for their effects on the formation parameters
(resistivity profile). However, without running resistivity tool's with similar depths and
volumes of investigation as the porosity tools, the resistivity profile may be useless for
interpreting and correcting porosity logs. In fact, the resistivity logs may not suggest
that invasion had occurred to any extent whilst the porosity log may suggest otherwise.
Woodhouse et al. (1991) examined the effects of gravity and mud density
contrast in driving vertical mud invasion in horizontal wells. Oil based mud (OBM)
falls in gas reservoirs, rises in water reservoirs and does not invade in oil reservoirs
[Figure 3.29]. In the case of OBM in a gas reservoir, initially gas saturation, Sg, is
reduced near the borehole. Sg increases once mudcake forms and the invaded ‘mud tear
drop’ will fall away leaving a trail of irreducible mud filtrate in the formation. LWD,
MAD and wireline logs can be used to document the dynamic formation fluid changes
[Figure 3.27]. Invasion is driven by the mud/formation pressure gradient, permeability
of the formation and mudcake, viscosity, diffusion, capillary pressure, gravity
segregation and porosity (Woodhouse et al. 1991). There are three steps to invasion:

• Spurt loss occurs before mudcake is formed, lasting only a few seconds.
• Dynamic loss occurs while drilling and during mud circulation. Mud circulation
erodes mudcake, allowing invasion to continue. When the rate of erosion equals the
rate of deposition, then Darcy’s Law governs fluid loss. Generally after 10-15 hours
of circulation, the rate of fluid loss is constant.
• Static loss occurs after mud circulation has stopped and is controlled by mudcake
permeability and not formation permeability. The volume of fluid loss is
proportional to the square root of time.

LWD nuclear tools are normally 150ft (45m) to 250ft (76m) behind the bit when
drilling a horizontal well but may be only 17ft (5m) behind the bit in vertical wells. At a
typical drilling speed of 12m/hr (39ft/hr), LWD nuclear tools are 6 hours (72m) behind
the bit. The invasion front may be beyond the depth of investigation of some tools.

• Neutron porosity values measured in multiple passes over several days in a gas well
can be seen to decrease (Woodhouse et al. 1991).

3-49
 Chapter 3: Horizontal well porosity anomalies

Gas Leg (Day 3) Gas Leg (Day 7)

I n c r e a s in g
time and depth

Smearing of
OBM ta k e s
p la c e

P iston lik e
displacement M±dle o il tear
of filtr a te drop descends

Figure 3.27 Under dynamic filtration, invasion in the oil leg is circular. In the gas
column, gravitational effects distort the invaded mud into an elliptical form. Under
equilibrium conditions this body drips free from the borehole and falls downwards
leaving a trail of residual oil redrawn (Figure 11 Woodhouse et al. 1991).

invasion

invasion

Invaded o il mud Invaded o i l

at residual o il descending off
s a t u r a t io n s h a le ed ge

Figure 3.28 Mud invasion profile versus time along a horizontal gas well redrawn
(Figure 12 Woodhouse et al. 1991).

3-50
 Chapter 3: Horizontal well porosity anomalies

Isotropic K Anisotropic K

Light Mud Heavy Mud

Figure 3.29 Permeability and mud density effect on invasion profiles redrawn (Figure
11 and 12 Cuddy et al. 1994).

In a horizontal gas well apparent porosity increased as OBM/WBM invaded

with a Hydrogen index (HI) of approximately 1.0 (Woodhouse et al. 1991). Once
mudcake had formed, gas (HI approximately 0) reinvaded the formation close to the
borehole and a lower apparent porosity was measured [Figure 3.28](Woodhouse et al.
1991). “No reasonable scenario could be proposed where gas would immiscibly work
around the OBM filtrate for the suggested filtrate rates and volumes, antithetic to liquid
phase pressure gradient and filtrate recharge flow” (Woodhouse et al. 1991).

3-51
 Chapter 3: Horizontal well porosity anomalies

If the neutron porosity interpretation were correct then other factors must affect
OBM filtrate behaviour. Kh/Kv>1000 is required for a non-circular invasion profile in a
horizontal well (Woodhouse et al. 1991). However, Kh/Kv rarely exceeds 10, so non­
circular invasion is unlikely. If Kh/Kv=l, OBM in gas would drop 12.9ft (3.93m) in 8
days after mudcake sealed the borehole (Woodhouse et al. 1991). In water OBM would
rise approximately 6ft (1.83m) in the same period of 8 days. If invasion were greater
than lOinches after 5-7 hours, the invasion front would be well beyond the depth of
investigation of porosity tools. The invaded zone measured by the tools would contain
OBM filtrate, connate water and residual gas.
Calculations show that gravity segregation occurs when K h /K v< 1 0 0 [Figure
3.27](Woodhouse et al. 1 9 9 1 ). In a vertical well the fluid is pushed away from the
borehole, but in a horizontal well the fluid is pushed around the borehole.

• Field evidence shows that when K h /K v< 1 0 , tear drops can form when dynamic
filtration is low (<2ml/hr-in2)(Woodhouse et al. 1 9 9 1 ).

Invasion estimates

To estimate the dynamic flow of invading mud filtrate at least two passes are
required (Harris et al. 1993). Resistivity can be used to estimate depth of invasion and
formation permeability may be calculated. The following equations do not account for
flow ahead of the bit, which may be a significant factor in highly permeable formations:

dV.mf 2xh(P„-Pf )
dt f \ (r \
Mmf Equation 3.1
In rf
rh M f
In
r
\ me Kf UJ
where Vmf= volume of invaded mud filtrate, t = time passed, h = formation thickness,
Pmf = mud filtrate viscosity, Kmc = mudcake permeability, juf = formation filtrate
viscosity, Kf = formation permeability, rb = borehole radius, rmc = distance to mudcake
from borehole centre, r/= distance to limit of reservoir, Pb = borehole fluid pressure and
P f- formation fluid pressure.

3-52
 Chapter 3: Horizontal well porosity anomalies

dt At li~ i\ 12 t j

where <j>= porosity, Sxo - water saturation of invaded zone, r,- = radius of invasion.
Assuming that no mudcake exists, at t=0, rmc=ri=rb,

/ f„ \ \
8^ L = 2nhrb^ =2 zh{Pb-P f )El In M -l Equation 3.3
Ot Ot 1=0 A/ V U J )

Combining Equation 3.2 and Equation 3.3 at t=0, rmc=ri=rb,

/ \
</Sxonhr‘ ^ =2xh{Pb- P f ) El In\ f 1-1 Equation 3.4
JUf V UJ

and rearrange for

2 {Pb - P f ]Kf t2
f / \
r,(ti ) = h + In '/ -1 Equation 3.5
V UJ )

A minimum amount of invasion is necessary to be able to detect a departure

from Rt and a difference between resistivity at two depths of investigation. Generally
invasion affects neutron porosity responses less than density responses because of the
greater depth of investigation of the neutron measurement (Allen et al. 1990; Jackson et
al. 1994). If invasion/alteration is less than linch then both LWD and wireline density
tools will detect the invasion and a non-zero DRHO will be recorded.

• The effect of mud invasion on the density values is dependent on density contrasts
and the depth of invasion.

3-53
 Chapter 3: Horizontal well porosity anomalies

Modelled invasion effects in density and neutron values

Non-uniform invasion (caused by permeability anisotropy) was investigated for

southern North Sea horizontal gas wells (Cowan and Wright 1997), because non-
uniform invasion was suspected to be the cause of anomalously high porosity readings
(Cuddy et al. 1994). Numerical modelling of wireline neutron porosity and density tools
for various sandstone porosity and saturation values in a horizontal well including the
effects of non-uniform invasion was carried out (Cowan and Wright 1997).

• The results indicated that the wireline density tool would read 6.5-19.0pu too high,
but the wireline neutron porosity tool would read 3-12.5pu too low in the non-
uniform case compared with the uniform case.

The numerical (Monte Carlo) modelling was laboratory benchmarked against

density (RODENT) and neutron tools (WILT), which have slightly shallower depths of
invasion (DOI) than commercial tools but display the same kind of differences observed
with commercial wireline tools. RODENT’s DOI is between 8-9cm whereas
commercial wireline tools are between ll-13cm. WILT’s DOI is between 20-26cm
whereas commercial wireline tools are between 24-30cm (Cowan and Wright 1997).
From Darcy’s Law [See Equation 1-4 Hilchie (1982)], a permeability ratio of
10:1 (horizontakvertical) would lead to invasion at the sides approximately three times
greater than above and below the borehole [Figure 3.30]. As invasion increases in gas
sand wireline density porosity decreases (increasing bulk density) and wireline neutron
porosity increases (increasing hydrogen index [HI]) as shown in Figure 3.31. The
density tool’s DOI increases slightly with porosity, but the neutron porosity tool’s DOI
decreases with porosity as the HI increases with porosity. There is a slight increase in
the neutron porosity tool’s DOI when the brine mud is replaced by oil based mud due to
the reduction in formation sigma (formation of absorption thermal neutrons).
Table 3.11 shows the maximum difference in apparent porosity between non-
uniform and uniform invasion at the respective uniform invasion depths at which they
occur. As the depth of non-uniform invasion increases density porosity increases and
neutron porosity decreases at a given equivalent uniform invasion depth [Figure 3.32].

3-54
 Chapter 3: Horizontal well porosity anomalies

Tool

Invaded Zon

Formation

Uniform Invasion

Non-Uniform Invasion
Aspect Ratio 3:1

Figure 3.30 Models of uniform and non-uniform invasion (Figure 2 Cowan and Wright
1997).

3-55
 Chapter 3: Horizontal well porosity anomalies

g . 10
a.

0 5 10 15 20 25 30 35 40 45 50
Invasion Depth (cm)
Apparent Density Porosity
——— Apparent Neutron Porosity

Figure 3.31 Apparent porosity against invasion depth for uniform brine invasion of a
20pu gas sandstone, lOsu water saturation (Figure 3 Cowan and Wright 1997).

o 18

14

0*

Q .10
-12
-14
0 5 10 15 20 25 30 35 40 45 50
Uniform Invasion Depth (cm)
Difference in Density Porosity
Difference in Neutron Porosity

Figure 3.32 Difference in apparent porosity between non-uniform and uniform brine
invasion against invasion depth for uniform brine invasion of a 20pu gas sandstone,
lOsu water saturation (Figure 4 Cowan and Wright 1997).

3-56
 Chapter 3: Horizontal well porosity anomalies

Formation Formation Borehole & Max Difference Apparent Density Max Difference Apparent Neutron
Porosity Saturation Invasion Porosity Porosity

(pu) (su) Fluid Difference (pu) Invasion (cm) Difference (pu) Invasion (cm)
10 10 Brine 6.5 10.0 -4.5 30.0
20 10 Brine 11.5 10.0 -9.0 18.0
20 10 Oil Mud 11.0 10.0 -6.0 24.0
20 30 Oil Mud 6.5 10.0 -3.0 30.0
30 10 Brine 18.0 8.0 -12.0 18.0

Table 3.11 Maximum differences in apparent porosity between non-uniform and

uniform Invasions (Table 3 Cowan and Wright 1997).

Tool Increase porosity Increase invading fluid density Increase water saturation
Density Increases Increases decreases
Neutron Increases Increases decreases

Table 3.12 Change in porosity difference when one parameter is increased, while the
other two remain constant (From the conclusions Cowan and Wright 1997).

The tools measured more gas-filled formation below the borehole than at the
sides because invasion was three times greater at the sides than below where the tools
measure. Density porosity values are greater but the neutron porosity values are lower
than expected, assuming uniform invasion. Note: full invasion for the density tool is
approximately 10cm, but 18-30cm for the neutron. Table 3.12 summaries the changes
in the density and neutron porosity difference between the uniform to non-uniform.
Sensitivity analyses of the density and neutron tools are summarised in Figure
3.33 for the density tool and Figure 3.34 for the neutron tool. The results show that the
density tool was highly forward focused with the long spaced detector three times more
sensitive to the formation than the short spaced detector and both detectors were
insensitive to the borehole fluid.

• The highly forward focused nature of the density tool means that the readings were
greatly affected by non-uniform invasion (Cowan and Wright 1997).
• The results show that the wireline neutron tool was less forward focused than the
wireline density tool, but it was still affected by non-uniform invasion (Cowan and
Wright 1997).

3-57
 Chapter 3: Horizontal well porosity anomalies

■135deg Long Spaced Detector 135deg

•\ ; s \ /
\

/ i ' n
r t '' V
1i i '
7 ' * \
19
|% I
I, 82% \
t V
-45deg -22.5deg -7.5deg 7.5deg 22.5deg 45deg

-135deg Short Spaced Detector 135deg

' / \ \
/ / i \
/ / -2%i i-2% X s
/ / 104% A v
/ I ^ ^ V \

-45deg -22.5deg -7.5deg 7.5deg 22.5deg 45deg

Figure 3.33 RODENT (density) sensitivity results (Figure 5 Cowan and Wright 1997).

3-58
 Chapter 3: Horizontal well porosity anomalies

-135deg 135deg

s 1%

V
\ /

13%
(3 V
13%

/ 73%

-45deg 45deg
Far Detector

-135deg 135deg

/
\

\ m

, 8 %! 8%

\
\
\
■
/ 84% ;X
fN H m M \ :
-45deg 45deg
Near Detector

Figure 3.34 WILT (neutron) sensitivity results (Figure 6 Cowan and Wright 1997).

3-59
 Chapter 3: Horizontal well porosity anomalies

Fractures

Modelling of the presence of fractures demonstrated' that the neutron tool has
some sensitivity to the sides of the borehole in a horizontal well, where fractures are
most likely to occur because of overburden pressure (Cowan and Wright 1997). The
maximum increase in porosity caused by fractures at the borehole sides would be lpu,
which would be insufficient to explain the anomalous neutron porosity values observed
in the southern North Sea (Cuddy et al. 1994). Further modelling of fractures at the side
of the borehole demonstrated that for a side facing density tool, compared with a down
facing density tool, the observed porosity could increase by 1 to 15pu. Observed
southern North Sea horizontal well porosity values between a down and side facing
density tool were approximately 3pu, so fractures would explain these observations
although fracture type may alter results. The rugosity was known from the scanner data
in this case.

• From the author’s recent employment borehole rugosity was a possible explanation
for the poor match of the neutron and density measurement with modelled results.

Effect o f porosity algorithm

The final part of the Cowan and Wright (1997) study involved a comparison of a
number of standard porosity calculations [Table 3.13, Table 3.14, Table 3.15, Figure
3.36 and Figure 3.37]. All the methods [see section 2.4.3] compared rely on density and
neutron porosity values alone, but the Shallow iterative and Wiley & Patchett (Wiley
and Patchett 1994) methods still overestimated porosity by 2 to 5pu. For southern North
Sea the observed porosity anomalies were between 2 to 4pu compared with core
porosity values (Cuddy et al. 1994).
Both Data Sets examined in this thesis [Chapters 4] were used to assess the
applicability of the above porosity algorithms on test data. Some success was evident
with Data Set 1 [Chapter 4] from the southern North Sea, but little success was gained
with application to Data Set 2 [See Chapter 4] due to the effects of clay.

3-60
 Chapter 3: Horizontal well porosity anomalies

.TD: BO /257
TO:8 6 /2 5 6

.TO: 6 0 /7 8

. 7 8 :4 1 / 3 2 4
"TO: 6 3 /2 6 3

. 7 0 :6 8 / 2 4 3

. 7 8 :7 5 /2 5 3
. 7 0 :5 6 /2 6 7

.7 8 :8 1 /7 8
.7 8 :8 7 / 2 5 9
.7 0 :6 9 /2 6 0

.7 8 :7 6 / 2 6 6
.7 8 .-7 6 /2 7 7

Figure 3.35 Ultrasonic images showing fractures at the sides of a horizontal borehole
(Figure 8 Bedford et al. 1997).

0 5 10 15 20 25 30 35 40 45 50
Invasion Depth (cm)
Square Root Method
Wiley & Patchett Method
Shallow Iterative Method

Figure 3.36 Predicted formation porosity against invasion depth for uniform brine
invasion for a 20pu gas sandstone at lOsu water saturation (Figure 7 Cowan and Wright
1997).

3-61
 Chapter 3: Horizontal well porosity anomalies

9
Ou

09
bO*
o
a*

*8
£
e
a>

,o>

15 20 25 30 35
Uniform Invasion Depth (cm)
Square Root Method
Wiley & Patchett Method
Shallow Iteration Method

Figure 3.37 Difference in predicted formation porosity between non-uniform and

uniform brine invasion for a 20pu gas sandstone at lOsu water saturation (Figure 8
Cowan and Wright 1997).

Square Wiley & Shallow Shallow

Formation Borehole
Formation Root Patchett Iterative Iterative
Saturation & Invasion Invasion
Porosity (pu) Porosity Porosity Porosity Sw Sxo
(su) Fluid
(P«) (pu) (PU) (su) (su)
Uniform 1.50 1.01 0.95 0.98 9.77
10 10 Brine
Non- Uniform 1.67 0.98 1.18 1.13 11.28
Uniform 2.97 1.83 1.95 1.05 10.53
20 10 Brine
Non- Uniform 3.05 1.59 2.09 1.02 10.18
Uniform 2.23 1.64 2.65 173 6.38
20 10 Oil Mud
Non- Uniform 2.23 2.26 1.87 1.14 4.23
Uniform 1.53 2.12 1.75 3.28 4.04
20 30 Oil Mud
Non- Uniform 1.40 2.53 1.14 2.03 2.50
Uniform 3.96 2.65 2.70 0.91 9.14
30 10 Brine
Non- Uniform 4.56 3.06 3.09 0.97 9.69
Average 2.51 1.97 1.94 1.42 7.77

Table 3.13 RMS values of porosity and saturation predictions (Table 5 Cowan and
Wright 1997).

3-62
 Chapter 3: Horizontal well porosity anomalies

Formation Formation Borehole & Square Root Wiley & Patchett Shallow Iteration
Porosity (pu) Saturation (su) Invasion Fluid Method (pu) Method (pu) Method (pu)

10 10 Brine 4.5 3.0 3.0

20 10 Brine 7.0 5.0 5.0
20 10 Oil Mud 6.0 5.0 5.0
20 30 Oil Mud 3.0 2.6 2.3
30 10 Brine 11.5 8.0 7.5

Table 3.14 Maximum predicted porosity differences (Table 6 Cowan and Wright
1997).
Saturation (su)
Formation Formation Borehole &
Invasion Porosity (pu)
Porosity (pu) Saturation (su) Invasion Fluid Sw Sxo

Uniform 1.07 1.01 10.09

10 10 Brine
Non-Uniform 1.12 1.04 10.44
Uniform 1.57 0.82 8.20
20 10 Brine
Non-Uniform 1.54 0.75 7.48
Uniform 1.51 0.90 3.31
20 10 Oil Mud
Non-Uniform 1.18 0.62 2.28
Uniform 1.03 1.66 2.05
20 30 Oil Mud
Non-Uniform 0.97 1.48 1.83
Uniform 2.46 0.81 8.09
30 10 Brine
Non-Uniform 2.90 0.93 9.26
Average 1.54 1.00 6.30

Table 3.15 RMS values of porosity and saturation for shallow iterative method with
Wiley and Patchett porosity estimates (Table 7 Cowan and Wright 1997).

• The square root method was shown to be the least reliable, while the Shallow
Iterative and Wiley & Patchett (Wiley and Patchett 1994) methods were comparable
and improved porosity estimation (Cowan and Wright 1997).

The sonic tool derived porosity best matched the expected porosity values from
offset vertical wells [Figure 3.4](Bedford et al. 1997). However, if fracturing and
washout is evident then the sonic porosity estimates will also be inaccurate. Running
sonic logs in horizontal wells is considered risky due to tool losses. LWD sonic tools
are available although their quality is a relative unknown. In horizontal wells the
recorded sonic velocity may only be representative of faster layers approximately
parallel to the borehole (Cuddy et al. 1994), also sonic anisotropy in shale layers could
affect values if shale beds are close enough. Data set 2 [See Chapter 4] demonstrates an
interesting variation in which the sandstone beds appear to be reading ‘fast’.

3-63
 Chapter 3: Horizontal well porosity anomalies

Most significant effects of invasion on the neutron measurement are observed

when the formation fluid is predominantly gas because neutrons are “gas seeking” (Ellis
and Chiaramonte 2000). When a neutron tool enters a water”zone from a gas zone, the
neutron tool only detects the water zone once the tool is within the water zone.
However, a neutron tool entering a gas zone from a water zone will detect the gas zone
from approximately 20cm (approximately 8inches) true vertical depth above Figure
3.38. Gas will lower measured neutron porosity.

• Neutron count rate analysis has revealed the presence of gas cap from within an oil
reservoir although the effect on the neutron porosity logs was very subtle
(Samworth 2000).

In a 40° deviated gas well the LWD neutron porosity was almost constant but
the density showed variation (Cannon 1998). The well was logged with azimuthal LWD
density, neutron and resistivity tools giving improved accuracy in the calculation of
porosity, Sw and mineralogy primarily due to depth of investigation, matching two of
the azimuthal resistivity measurements with the density (Shallow) and neutron (Deep)
measurement [Figure 3.39]. In addition, direct measurement of the borehole diameter
with ultrasonic transducers has improved the quality of the density and neutron
measurements. In this case the azimuthal LWD density and neutron porosity tool
recorded lOpu density variations between bottom, left and right sectors, which
corresponded with LWD shallow resistivity images. The mud filtrate was more
conductive and denser than the gas filled formation [Figure 3.41](Cannon and Kienitz
1999). The neutron porosity measurement tended to see only gas, but the density
measurement recorded varying amounts of mud invasion. The mud invasion was deeper
underneath the borehole. The density values increased in gas compared with the ‘true’
density (water filled) because the mud appeared to be heavy to the density measurement
i.e. a negative DRHO was produced as the tool tried to remove the effect (Cannon and
Kienitz 1999). The differential invasion causes the negative DRHO, resulting in density
porosity estimates 5-6pu too high but overestimates of lOpu were recorded.

• Teardrop invasion can be identified in an in-gauge horizontal gas well with

resistivity and density images [Figure 3.41](Cannon and Kienitz 1999).

3-64
 Chapter 3: Horizontal well porosity anomalies

25 PU Sand

boreho
Gas Sand

25

20
3
CL

c/3
O
j—
o
CL
0
c
o
CO
0
E Top
-4—»
c
0
03
Q.
Q.
<
ottom

- 50 4 0 3 0 20-10 0 10 20 30 40
- - -

Distance to top of gas bed (cm)

Figure 3.38 Chart to show the effect on the modelled response of a LWD azimuthal
neutron porosity tool in a horizontal well (89°) passing through a 25pu water filled sand
into a 25pu gas filled sand. The tool provides four neutron porosity measurements from
which the TOP detector faces ‘up’ and the bottom detector faces ‘down’, redrawn from
(Slide 22 Ellis and Chiaramonte 2000). Note: Values in standard apparent limestone
porosity units.

3-65
 Chapter 3: Horizontal well porosity anomalies

RAB G -F ac to r with
D ensity/N eutron
Rt/Rxo = 50

— Density
— «*r- Neutron
Shallow.
Medium
Deep
— Ring
vtMttxm 0 jt

0 5 10 15 20 25 30
R a d i u s from weilbore
R i( in.)
Figure 3.39 LWD density and neutron responses compared with resistivity at bit (RAB)
responses (Figure 6 Cannon and Kienitz 1999). Note: The close match of the shallow
curve and density, and the deep curve and neutron in this environment.

25 p.u. Gas Sand in 8-in. Borehole

12. , -------

Bottom

Filtrate

L, "Teardrop" Length (cm)

Figure 3.40 Modelled horizontal gas well LWD neutron responses sand with tear drop
mud invasion (Figure 5 Cannon and Kienitz 1999). An invasion difference of 5cm in
the shallow direction has a bigger effect than a 20cm difference in the draped direction,
confirming that the gas closest to the well dominates the neutron response. Note: min
is the minimum invasion length and L is the maximum (teardrop or draped invasion).

3-66
 Chapter 3: Horizontal well porosity anomalies

RAB Resistivity Porosity

RAB Vision
Shallow Button Down Density Up Resistivity Density
M '- L - l 1 i i \.m
Shallow Button Up Density Down

Deep Button Up Neutron

ohgt-m____

m
■
m
m
*

m
•

m
- -

u m m

Figure 3.41 The left-hand column contains from left to right the RAB shallow button
down, shallow button up, deep button up and CDR (combinable dual resistivity)
attenuation. The darker shading indicates additional invasion on the bottom of the hole
that is not detected on the top. In track 2 from left to right are density up, density down
and neutron. The shading between the two density curves also indicates additional
invasion down. There are two images on the right from RAB and ADN tools. Each
image is presented with the top of the hole on the left, bottom in the middle and back to
the top on the right. The darker the RAB image the more filtrate present. The lighter the
ADN image the more filtrate present. Even though the ADN tool does not have the
same resolution as the RAB tool, it still finds the maximum invasion in the same
direction as the RAB tool, which is not always down (Figure 10 Cannon and Kienitz
1999).

3-67
 Chapter 3: Horizontal well porosity anomalies

As a rule: “All the filtrate that leaves the well eventually migrates downdip,
even the filtrate that leaves from the updip side of the wellbore” (Cannon and Kienitz
1999). This effect is accelerated if vertical permeability is large and can lead to an
apparent water leg when filtrate ponds above the impermeable layer. The neutron
measurement is dominated by the gas closest to the tool almost regardless of azimuth
[Figure 3.40]. The neutron log appears to be ‘lazy’ (varies only slowly) and appears to
ignore the invasion. However, the new LWD tools do not solve all the problems
associated with horizontal wells (Cannon and Kienitz 1999). Fluid distributions, well
deviation and bed dip require 3D modelling due to asymmetric mud invasion, which
require 3D log measurements. At present, 2D plus dip modelling is available and is
performed on each quadrant and the measurements are not yet fully 3D.

• The density measurements vary considerably in the presence of teardrop invasion,

the neutron measurement does not (Cannon and Kienitz 1999).
• The teardrop does not always form in the down direction but moves in the down
direction of the bedding [Figure 3.41](Cannon and Kienitz 1999).

Fluid density changes between the two logging runs in the same borehole cause
problems when comparing LWD and wireline density values (Hansen and Shray 1996).
Tools capable of measuring elliptical invasion profiles are required to understand fully
the borehole environment with associated software to model for the effects of borehole
signal, shoulder beds, non-circular invasion with fluid segregation, dipping beds and
anisotropy. Likewise, “The development of a computer program that simultaneously
inverts resistivity and porosity logs, by minimising the differences between real tool
readings and reconstructed logs is recommended” (Peeters et al. 1999).
Mud invasion is more complex than has been described above. Although a
piston invasion model may be adequate for interpretation of deep resistivity curves, the
model is not sufficiently detailed for near borehole invasion profile (Peeters et al.
1999). Saturation and salinity fronts exist as well as invasion fronts [Figure 3.42]. OBM
may alter wettability and reduce Sw below irreducible levels possibly dehydrating
shales. Clearly changes in formation properties occur in space and time (4D). An
additional problem with LWD logs is that the tools have far higher chance of loss of
formation contact compared with their wireline counterparts.

3-68
 Chapter 3: Horizontal well porosity anomalies

200

U ndisturbed
zo n e
F orm ation
w ater b ank
M ud filtrate
bank
\ Resistivity
Resistivity transition zo n e
transition zo n e • and
a n d salinity front sa tu ratio n front
02

0
Sw,% W ater Fluid
Type
100

ormaboo W ater
Salinity Mud filtrate Salinity

Radial D ista n c e From W ellbore

Figure 3.42 Two front invasion model of fluid displacement (Figure 3 Peeters et al.
1999).

Alteration

Formation alteration in the vicinity of the borehole generally affects neutron

porosity less than density measurement because of the larger depth of invasion of the
neutron measurement (Allen et al. 1990; Jackson et al. 1994). Both density (decreasing
bulk density) and neutron (increasing HI) porosity values will increase because of water
absorption from the drilling mud. Drilling induced fractures in a horizontal well are
likely to be at the sides of the borehole because the overburden pressure represents the
maximum principal stress. The partial collapse will further increase K h /K v and therefore
invasion at the sides of the borehole [Figure 3.35](Cuddy et al., Day et al. 1994).
Fracturing can affect DRHO if it is serious and the density caliper may indicate if break
out occurs along the roof of the borehole.

3-69
 Chapter 3: Horizontal well porosity anomalies

A B C D 10

", ' ■V
I _____________________
sand dtrully gamma ray aonlc m b tiv ity

■ (W 2-5 100 120 2-5

WELL DEVIATION - M' shale ) 'u n d ‘ 2.2 20 *0 50

Table 2 • Summary of log measurements

Figure 7a - Formation model used to simulate curves in Fig 7b • 7c. used for simulation In Figure
The upper and lower parts of the wellbore cut the bed 7b-c.
boundaries at measured depths A, B, C and D.
Log measurements in each formation are summarized
In Table 2.

Figure 7b.

Phase resistivity RPS Simulation of Induction

and CDR curves across
the 2’ sand shown In
Figure 7a.
Formation resistivity
Resistivity
(ohm m.)
Amplitude resistivity
RAD

A B C D Measured Dcptl

Formation resistivity

Medium Induction
Deep Induction

IS
Figure 7c
DENSITY
(gm/cc.) Simulation of density,
2.2 gamma ray and sonic
across the formation
100 shown in Figure 7a.
(Measured depths A, B
C and D correspond to
GAMMA RAY
those in Figure 7a.)
(GAPI)
20
120

SONIC
(ps/ft)
90

A B C D
SENSOR PLACED ON ENSOR PLACED ON
LOWER SIDE OF — s PPER SIDE OF
BOREHOLE BOREHOLE

Figure 3.43 Schematic to illustrate bed boundary detection in horizontal wells (Page 10
White 1991).

3-70
 Chapter 3: Horizontal well porosity anomalies

Downhole temperature will change between LWD and wireline tool runs. Clay
alteration between drilling, LWD, MAD and wireline runs can alter formation porosity
by 2-6pu because of water absorption from the mud (Bedford et al. 1997). This process
reduces formation density, increases the hydrogen content of the formation and alters
the elastic properties, increasing all the porosity estimates.

• Inaccurate neutron and density porosity in shale will shift the shale point in the
neutron-density crossplot (Allen et al. 1990; Jackson et al. 1994).

Bedding

In horizontal wells the orientations of log measuring systems are usually

anisotropic with respect to bedding (Bigelow et al. 1992). An unusual effect was
observed in a numerically modelled well with a deviation of 89°. The borehole was
intersected by a 6inch thick low porosity horizontal bed (90°), producing density and
neutron responses such that an apparent gas crossover was bordered by apparent shale
effect (Ellis and Chiaramonte 2000).

• The depth of investigation of the tools (LWD and wireline) does affect the position
along the borehole at which bed boundaries are detected (White 1991; Singer 1992).

Different tools will measure different lithologies and/or fluids simultaneously at

the same depth location [Figure 3.5 and Figure 3.43](White 1991; Singer 1992).

3-71
 Chapter 3: Horizontal well porosity anomalies

Summary o f alteredformation conditions

Altered formation conditions often degrade the quality of porosity logs. The
effect of drilling the borehole and the mud used are to alter the formation from its native
state. Mud invasion replaces the native fluids and may alter clay minerals by hydration
or dehydration. Drilling can induce fracturing in the near borehole environment. Both
the invasion and fracturing processes are dependent on the orientation of the borehole
with respect to the applied forces (e.g. gravitational and regional stress fields).

Invasion

• Dynamic loss occurs while drilling and during mud circulation. Mud circulation
erodes mudcake, allowing invasion to continue. When the rate of erosion equals the
rate of deposition, then Darcy’s Law governs fluid loss (Woodhouse et al. 1991).
• Static loss occurs after mud circulation has stopped and is controlled by mudcake
permeability and not formation permeability. The volume of fluid loss is
proportional to the square root of time (Woodhouse et al. 1991).
• Generally the invasion process is static by the time wireline tools are logged (>8 to
lOhrs after bit penetration, equilibrium is established)(Woodhouse et al. 1991).
• Gravity and permeability anisotropy have a combined effect on mud invasion in
horizontal wells that may produce non-uniform invasion profiles, consequently
affecting log measurements (Woodhouse et al. 1991).
• Field evidence shows that when K h /K v< 1 0 , tear drops can form when dynamic
filtration is low (<2ml/hr-in2)(Woodhouse et al. 1991).
• Teardrop invasion can be identified in an in-gauge horizontal gas well with
resistivity and density images. The teardrop moves in the down direction of the
bedding [Figure 3.41](Cannon and Kienitz 1999).

Effect on logs of altered formation conditions

• Neutron porosity values measured in a gas well with multiple passes over several
days decrease as gas migrates towards the borehole (Woodhouse et al. 1991)

3-72
 Chapter 3: Horizontal well porosity anomalies

• The effect of mud invasion on the density values is dependent on density contrasts
and the depth of invasion (Woodhouse et al. 1991).
• When invasion is 3 times greater at the sides than above and below the well
compared with uniform invasion, modelling indicated that the wireline density tool
would read 6.5 to 19.0pu too high, but the wireline neutron porosity tool would read
3 to 12.5pu too low (Cowan and Wright 1997).
• The square root method was shown to be the least reliable, while the Shallow
iterative and Wiley & Patchett (Wiley and Patchett 1994) methods were comparable
and improved porosity estimation (Cowan and Wright 1997).
• The density measurements vary considerably in the presence of teardrop invasion,
the neutron measurement does not (Cannon and Kienitz 1999).
• Inaccurate neutron and density porosity in shale will shift the shale point in the
neutron-density crossplot (Allen et al. 1990; Jackson et al. 1994).
• The depth of investigation of the tools (LWD and wireline) does affect the position
along the borehole of the detection of bed boundaries (White 1991; Singer 1992).

The author’s experience of altered formation conditions is that the effects of

invasion are common in all wells and that fractures are often only identifiable from
image logs. Clay alteration is frequently encountered and can cause the well to be
bridged preventing wireline tool access beyond that point without special techniques. In
horizontal wells formation alteration and features may be overlooked as they only
change very gradually along the length of the well. Plotting the logs on a reduced scale
such as 1:1000 can help identify unusual features like increased depth of invasion due
to a change in permeability.

• Neutron count rate analysis has revealed the presence of a gas cap from within an
oil reservoir although the effect on the neutron porosity logs was very subtle
(Samworth 2000).

3-73
 Chapter 3: Horizontal well porosity anomalies

3.4 Summary

This chapter provided published evidence of anomalous porosity values in

horizontal wells. Two North Sea studies (Cuddy et al. 1994; Austin et al. 1994)
concluded that horizontal well porosity values were too high compared with expected
values from vertical wells. In addition an overview of the complications experienced
when evaluating horizontal well porosity logs was given, perturbations include:

• Tool effects. Preferential tool orientation, differences in sensor physics, differences

in source-detector spacings, investigation depths, spectral responses, method of
source-detector to formation contact, basic reference standards differences, and
environmental corrections to common standard conditions.
• Borehole conditions. Debris, fractured/unstable borehole, hole size, mud density
changes and additives (Barite, KC1 and Hematite), standoff and rugosity.
• Altered formation conditions. Mud invasion, bed boundary dip and clay alteration.

Jackson et al. (1994) provides a helpful discussion of the effects of differences

of LWD and wireline values on formation evaluation from tool errors, borehole and
altered formation conditions. Allen et al. (1990) provides a guide for the porosity
logging method to be employed dependent on borehole conditions expected.

Porosity anomalies

Cuddy et al. (1994) compared a large number of vertical and horizontal gas
wells from the southern North Sea and illustrated that log porosity values overestimated
porosity in horizontal wells regardless of tool type [Figure 3.3 and Table 3.1] and the
overestimates appeared to increase with increased porosity and permeability.
Bedford et al. (1997) demonstrated that in one horizontal well, downward facing
wireline density tool porosity values measured higher porosity compared with vertical
wells in that particular field due to the additional gas effect on density porosity from
anisotropic invasion. Side facing density porosity values unexpectedly measured greater
porosity values than the downward facing density values. Ultrasonic images revealed
fractures in the side of the borehole that increased the formation porosity calculated.

3-74
 Chapter 3: Horizontal well porosity anomalies

Sonic porosity values were comparable with the vertical well values. The solution was
to gas correct the down facing density porosity values using the calculated water
saturation values thus obtain a good match with the sonic porosity values.
In a detailed study (Austin et al. 1994) of southern North Sea acoustic log
responses, sonic travel time values were 2-3 ps/ft lower than core porosity values at 6pu
porosity, but approximately 10ps/ft lower than expected at 15pu porosity. This would
have led to overestimates of porosity by 5pu and underestimates of rock strength,
resulting in conservative well completion strategies. Image logs were run to assess
situations where ‘fast’ beds were influencing the sonic logs. The acoustic core
measurements and theoretical calculations were unable to account for the anomalous
sonic porosity values, either by the presence of gas or in-situ stresses.

In the few horizontal wells in which sonic tools are run (approximately 5 to
10%), sonic porosity seldom matches the other calculated porosities (Samworth 2000).
This is especially so in low porosity formations (<5pu) in which the sonic travel time
appears to be independent of porosity. In addition, borehole conditions (fractures) and
the proximity of ‘fast’ beds may affect sonic porosity values, in which case density
porosity values would be used. Sonic tools are rarely run in horizontal wells due to the
sonic tool’s physical weakness and thus the increased risk of toolstring loss.

Tool errors

Differences in detailed tool design and toolstring configuration lead to LWD

and wireline measurements that do not always replicate each others responses. Common
standards for all logging tools need to be established in a wider range of lithologies than
currently available. Poor understanding and control of logging tool position within the
borehole whilst logging leads to poor porosity logs particularly in horizontal wells and
wells in which lithologies that deviate significantly from calibration points are
encountered.

• Wireline logs from tools built by different companies can vary significantly as
conditions deviate from the standard conditions (Coope 1983; Sakurai et al 1992).

3-75
 Chapter 3: Horizontal well porosity anomalies

• Differences in sensor physics can produce different wireline and LWD logs in
identical conditions due to the slightly different source-detector spacings, depths of
investigation and spectral responses (Jackson et al. 1994; Sakurai et al. 1992).
• Method of source-detector to formation contact will affect responses (Jackson et al.
1994; Sakurai et al. 1992).
• Basic reference standards are different (Jackson et al. 1994).
• LWD neutron porosity values in horizontal wells should measure lower values than
the same tool would in an equivalent vertical well (Day and Petler 1990).
• LWD tools should be stabilised with full gauge stabilisers immediately above and
below the tool to reduce lateral drillstring motion and poor centralisation, which
would lead to erratic readings (Cunningham et al. 1990).

The author’s opinion is that the problems of logging tool deployment (either
LWD or wireline) in horizontal wells is not sufficiently addressed in published papers.
For instance wireline tools may not be running in hole as intended facing the floor
without rocking to the sides and that inclination and deviation data should routinely be
recorded for all horizontal wells (including LWD) as log quality control. Common
calibration standards for wireline and LWD tools in limestone, dolomite, sandstone and
importantly shale need to be established. The author’s experience of LWD logs are they
are often poor quality when compared with wireline data, the main differences being
depth of investigation, bed boundary resolution and noise rejection of the LWD data.
Other problems occur (Samworth 2000) such as:

• Non-parallel standoff from LWD and/or wireline tool running across a chord could
result in the LWD and wireline differences.
• Time lapse density porosity from up and down wireline logs in which several hours
have elapsed between passes do not appear to show any discernible differences.
• Running two orthogonal density tools can lead to the two density tools pulling
towards one another giving two poor density logs.

3-76
 Chapter 3: Horizontal well porosity anomalies

Borehole conditions

Poor borehole conditions can result in poor logs. Problems affecting horizontal
porosity logs: worse conditions than a comparable vertical well due to overburden
driven breakout; debris collecting along the length of the well and the composition of
the drilling mud. The selection of logging tools should be based on the expected
borehole conditions (Allen et al. 1990), some of the problems are:

• LWD density tools in sliding mode, the standoff algorithm may become ineffective
because of the dependence on the short spaced detector values from all azimuths
(Allen et al. 1990; Best et al. 1990).
• Debris in horizontal wells collects along the borehole floor and can introduce
variable standoff (Woodhouse et al. 1991).
• The spiral rugosity periodicity is twice the bit to first stabiliser distance and the
modulation depends on the weight on bit and bent sub angle (Betts et al. 1990).
• For a 40pu formation and a 2inch standoff, the error for a wireline neutron tool is
approximately 1.5 to 3.0pu, but approximately 12pu for a LWD tool (Allen et al.
1990).
• The LWD porosity tools performed best in OBM, followed by NaCl mud and were
poorer still in KC1 mud (Norve et al. 1989).

The author’s experience is that often LWD logs are poor quality when compared
with wireline because borehole degradation occurs prior to LWD logging therefore
there is little difference in borehole conditions, wireline porosity tools normally cope
better with the borehole condition than LWD. However, in some areas of the World
(e.g. The Caspian), where washout and the risk of borehole collapse are significant,
wireline tools are seldom used. Other problems occur such as:

• Numerous wells with mud weights well below 10.51bm/gal (1.26g/cm3) especially
OBM in which barite can effect PEF logs to the point that they are rendered useless.
• Wireline tool rollover leads to bad formation contact resulting in poor density
values indicated under-gauge caliper values due to tool weight partially collapsing
the caliper and possibly non-zero DRHO (Samworth 2000).

3-77
 Chapter 3: Horizontal well porosity anomalies

• BHA motion is not strictly rugosity but may manifest itself as variable standoff.

Altered formation conditions

Degraded porosity logs result from mud invasion replacing the native fluids and
altering clay minerals by hydration or dehydration. Invasion and formation alteration
can affect log values by:

• Gravity and permeability anisotropy in horizontal wells combined effect on mud

invasion may produce non-uniform invasion profiles (Woodhouse et al. 1991).
Teardrop invasion moves in the down direction of the bedding [Figure 3.41]
(Cannon and Kienitz 1999).
• Neutron porosity values measured in a gas well with multiple passes over several
days decrease as gas migrates towards the borehole (Woodhouse et al. 1991)
• Density values are dependent on mud density contrasts and invasion depth
(Woodhouse et al. 1991). Density values vary considerably in the presence of
teardrop invasion, the neutron values do not (Cannon and Kienitz 1999).
• When invasion is 3 times greater at the sides than above and below the well
compared with uniform invasion modelling indicates that the wireline density tool
would read 6.5 to 19.0pu too high, but the wireline neutron porosity tool would read
3 to 12.5pu too low (Cowan and Wright 1997).
• In horizontal wells in the presence of invasion 3 times greater at the sides than
above and below the well, the square root method was shown to be the least
reliable, while the Shallow iterative and Wiley & Patchett (Wiley and Patchett
1994) methods were comparable and improved porosity estimation (Cowan and
Wright 1997).
• Inaccurate neutron and density porosity in shale will shift the shale point in the
neutron-density crossplot (Allen et al. 1990; Jackson et al. 1994).
• The depth of investigation of the tools (LWD and wireline) does affect the position
along the borehole of the detection of bed boundaries (White 1991; Singer 1992).
• Neutron count rate analysis can reveal a gas cap from within an oil reservoir
although the effect on the neutron porosity logs is very subtle (Samworth 2000).

3-78
 Chapter 3: Horizontal well porosity anomalies

The author’s experiences are that altered formation conditions are common in
all wells. Clay alteration is frequently encountered and can cause the borehole to be
bridged preventing wireline tool access beyond that point without special deployment
techniques. The effects of invasion and fractures are often only identifiable from image
logs. In horizontal wells formation alteration and features may be overlooked as they
only change very gradually along the length of the well and plotting the logs on a
reduced scale such as 1:1000 can help identify such features.

Horizontal well formation evaluation is still in its infancy. This is not aided by
the use of tools designed for vertical wells and the lack of appropriate software to
handle the complexity of these situations. In the future tools designed for horizontal
wells are required that will provide azimuthal images at several depths of investigation.
Resistivity tools are needed with depths of investigation matched to the porosity tools in
combination with the deep resistivity measurements are essential. These tools already
exist for LWD but in wireline mode. Multiple logging runs will be required to acquire
the time dependency of invasion related factors and the use of affordable and quick
four-dimensional interpretation software will enable complete formation evaluation.

3-79
 Chapter 4: Analysis of the Data Sets

Chapter 4: Analysis of the Data Sets

This chapter examines two data sets, one from a southern North Sea gas field
(Data Set 1) and the second from a northern North Sea oil and gas field (Data Set 2).
The southern North Sea example consists of data from one vertical well and one
horizontal sidetrack well that demonstrated inconsistencies in porosity values. The data
from the vertical well include wireline, LWD and core data, but only LWD data for the
horizontal sidetrack well. The horizontal well LWD porosity values (maximum density
and neutron) were lower than those calculated from the vertical well.
The northern North Sea oil and gas field example oil leg is being developed with
numerous horizontal wells [Figure 4.3] and an understanding of any unusual porosity
values would be advantageous. The data set comprises log data from five wells, two
vertical wells (well 1, 33° from vertical, and well 5, 21°), one horizontal sidetrack and
two horizontal wells (wells lz, 3 and 6, all 90°). Core data was available for both
vertical wells and one horizontal well (well lz). The horizontal well porosity values
from the three horizontal wells were significantly different from the measured values
from the vertical well data. The data demonstrate that there are no simple explanations
for horizontal well porosity anomalies. However, this data set highlights the need for
image logs in horizontal wells and the use of frequency analysis to aid log
interpretation.
This chapter aims to draw conclusions from the data analyses presented and to
identify the cause(s) of the observed porosity anomalies. Firstly, an explanation of the
data sets and of the horizontal well porosity anomalies observed are given. Porosity
analysis of the data sets includes the use of several different porosity algorithms
(§2.4.3) and splitting the reservoir into a number of generic units to assess the
applicability of the algorithms using statistical tests. Spectral analysis of the data sets
enables the evaluation of contributing factors and the removal of some these unwanted
influences. Finally, conclusions are presented that detail the causes of the porosity
inconsistencies.
Note that descriptions of the geology of data set 1 and data set are presented in
appendices 1 and 2 respectively.

4-1
 Chapter 4: Analysis o f the Data Sets

4.1 Description of data sets

4.1.1 Data set 1

The data set consists of logging data from a vertical well and its horizontal
sidetrack well [Figure 4.1 and Figure 4.2]. Both wells were drilled using 8.5inch bits
with barite weighted 10.5 lbm/gal (1.26 g/cm3) OBM (70% oil, 30% water). The
‘vertical’ well was deviated at 57° from vertical through the reservoir (§3.1). Both
wireline and LWD tools were run in the vertical well with a complete core section
through the reservoir. This was achieved by coring the well first, then logging with the
wireline tools 36 hours after mud circulation was stopped. The LWD tools were run 2
days after the wireline tools. The wireline tools included gamma ray, litho-density,
neutron porosity, acoustic travel time, induction resistivity and in a later run formation
tester [Figure 4.1]. The LWD tools run were gamma ray, litho-density, neutron porosity
(sandstone corrected porosity is used throughout this thesis unless otherwise stated),
and electromagnetic propagation resistivity [Figure 4.1]. The core data included
horizontal and vertical helium porosity (overburden corrected), vertical and horizontal
permeability, gamma ray and grain density [Table 4.2 and Table 4.10]. Horizontal core
plugs were cut either perpendicular to the long axis of the core section or parallel with
any visible bedding present. The vertical core plugs are cut perpendicular to the
horizontal core plugs. The horizontal sidetrack well was horizontal (88°) in the Leman
sands. Only LWD tools were run in the horizontal well [Figure 4.2].
The vertical well intersected 134ft (40m) measured depth of Weissliegend at the
top of the reservoir, 94 ft (28 m) of Rotliegend dune/interdune sands in the bottom of
the well [Figure 4.1]. The horizontal well intersected 772 ft (235 m) of Weissliegend at
the top of the reservoir, 749 ft(228 m) of Rotliegend dune/interdune sands in the middle
and 405 ft (123 m) of Weissliegend at end of the well [Figure 4.2], Geological
description of the field for data set 1 is given appendix 1.
Wireline and LWD logs were in generally good agreement in the vertical well as
discussed below [Figure 4.1]. The caliper logs indicated that the vertical well was in­
gauge and smooth. However, the horizontal well LWD caliper had increased on average
by (0.6±0.2) inches greater than the bit size, therefore the horizontal borehole
conditions were significantly more rugose and/or non-circular than the vertical well
[Table 4.1].

4-2
 Chapter 4: Analysis o f the Data Sets

The wallet at the back of the thesis contains all the log plots.

Figure 4.1 Data set 1 vertical well 1 wireline and LWD logs. Note, the apparently high
permeability values are due to the scale allowing “back-up” so these values are in fact
read between 0.02 to 0.2 mD, not 200 to 2000 mD.

The wallet at the back of the thesis contains all the log plots.

Figure 4.2 Data set 1 horizontal well lz LWD Horizontal well logs. Note, the periodic
signature in the log curves with a wavelength of ~8 ft.

4-3
 Chapter 4: Analysis o f the Data Sets

The LWD caliper measurement used is not independent of the density

measurement (§3.3.1)(Best et al. 1990). Any unusual values in the density measurement
will be reflected directly in the caliper values and therefore it is very difficult to identify
the source of the measured density perturbation, although simply incorrect calculation
of the apparent caliper can not be ruled out. Vertical well wireline density and neutron
porosity values were slightly lower than the LWD values [Figure 4.1 and Table 4.1].
Vertical well wireline acoustic and density derived porosity values matched closely
with the core porosity values [Figure 4.16].
The horizontal well LWD maximum density porosity and neutron porosity
values were lower than expected from the vertical well values [Figure 4.16a, Figure
4.18a and Table 4.2]. The LWD maximum density derived porosity values were in good
agreement with the vertical well core porosity values [Figure 4.18a]. This indicated that
the formation fluids were flushed from in front of the density tool’s detectors and there
was little gas to perturb the maximum density measurements. The horizontal well LWD
average bulk density values were 0.18g/cm3 lower (9pu greater) than the maximum
bulk density due to borehole enlargement and/or LWD tool eccentring.
The respective density corrections in the vertical well are small showing that
there was minimal mudcake thickness. The wireline density correction was slightly
positive which may indicate some presence of gas. The average LWD density
correction in the vertical well was -0.012g/cm3 resulting from barite in the mud, but
+0.020g/cm3 in the horizontal well due to standoff [Table 4.1].
The vertical well wireline and LWD photoelectric factor values were both
greater than expected for sandstone (1.745B/e), demonstrating that the barite (266.8B/e)
in the mud had effected the logs. The horizontal well LWD PEF was significantly
greater (+2.3B/e) than the vertical well responses [Table 4.1]. This demonstrated that
the standoff/mudcake must have been at least 0.15 inches thick (Carpenter et al. 1997).
However, this may be due to barite in the invaded mud perturbing the measurement.
This is supported by the fact the LWD PEF log follows the rate of penetration log
indicating mudcake build-up/mud invasion [Figure 4.2]. In an overlying anhydrite bed
(5.055B/e), the PEF was 6B/e in the vertical well and close in value for both the
wireline and LWD tools. However, the equivalent anhydrite bed measured >12B/e in
the horizontal well (well deviation >70°). The horizontal well PEF was responding to
lithology changes, but the values were far greater than expected due to the barite.

4-4
 Chapter 4: Analysis o f the Data Sets

V ariable M inimum M aximum M ean S tan d ard Deviation

Am plitude Resistivity (Qm) 2.068 159.519 9.698 13.621

Phase Resistivity (Qm) 1.890 76.408 9.290 9.297

Maximum Density (g/cm3) 2.278 2.683 2.435 0.059
Average Density (g/cm3) 2.277 2.659 2.424 0.058
Density Correction (g/cm3) -0.039 0.011 -0.012 0.009
Neutron Porosity (% ) 0.020 0.157 0.095 0.031
Gam m a Ray (API) 31.463 89.086 38.188 5.403
Photoelectric Factor (B/e) 2.447 3.225 2.694 0.096
Rate of Penetration (*5 ft/hr) 101.363 190.252 142.824 20.008
C aliper (inches) 8.500 8.656 8.536 0.036

Deep Resistivity (Qm) 2.237 281.081 12.650 23.073

M edium Resistivity (Qm) 2.198 53.409 8.831 7.239
Density (g/cm3) 2.319 2.632 2.466 0.058
Density Correction (g/cm3) -0.018 0.056 0.017 0.010
N eutron Porosity (% ) 0.013 0.144 0.081 0.034
Acoustic Travel Time (psec/ft) 58.687 83.493 69.195 5.959
G am m a Ray (API) 34.228 62.947 41.037 3.643
Photoelectric Factor (B/e) 0.656 4.269 3.036 0.339
C aliper (inches) 8.572 8.596 8.595 0.002

Resistivity (Qm)* 4.145 23.158 8.197 3.352

M aximum Density (g/cm3) 2.196 2.645 2.482 0.073
Average Density (g/cm3) 1.908 2.514 2.309 0.087
Density C orrection (g/cm3) -0.089 0.120 0.020 0.020
N eutron Porosity (% ) 0.019 0.176 0.088 0.024
G am m a Ray (API) 20.556 138.309 32.036 5.806
Photoelectric F actor (B/e) 2.705 21.761 4.995 1.663
R ate of P enetration (x5 ft/hr) 9.828 170.186 35.418 16.748
C aliper (inches) 8.629 10.236 9.122 0.266

Table 4.1 Log value statistics. * resistivity values were modified to provide a
continuous log for calculations.

4-5
 Chapter 4: Analysis o f the Data Sets

V ariable M inim um M aximum M ean S tan d ard Deviation

Average Density Porosity (% ) -0.005 0.226 0.137 0.035

M axim um Density Porosity (% ) -0.020 0.225 0.131 0.036
N eutron Porosity (% ) 0.020 0.157 0.095 0.031

H orizontal Helium Porosity (% ) 0.054 0.194 0.114 0.033

Horizontal Perm eability (mD) 0.020 88.000 5.180 13.380
V ertical Perm eability (mD) 0.010 44.000 1.860 5.310

Acoustic Porosity (% ) 0.028 0.213 0.106 0.044

Density Porosity (% ) 0.011 0.201 0.112 0.035
N eutron Porosity (% ) 0.013 0.144 0.081 0.034

Average Density Porosity (% ) -0.092 0.450 0.193 0.073

M aximum Density Porosity (% ) -0.249 0.275 0.085 0.078
Neutron Porosity (% ) -0.034 0.301 0.083 0.033

Table 4.2 Derived porosity statistics (fluid: lg/cm3 density and 189ps/ft travel time).

The vertical well wireline and LWD resistivity values [Figure 4.1 and Table 4.1]
were in good agreement. This suggested that the mud invasion profile was similar when
wireline and LWD tools were logged. The resistivity tools used all have a depth of
investigation greater than 10", beyond the investigation depth of the porosity tools.
The horizontal well amplitude resistivity measurement failed downhole and was
not present between depths xl31-x687ft and x961.5-x962.5ft. A linear regression was
used at these depths to obtain resistivity and between x l 878-1987ft where data
appeared to be spurious. The horizontal well phase resistivity log was continuous. A
linear regression between phase and amplitude resistivity provided a resistivity log for
continuous water saturation calculation in the horizontal well.
The core permeability values and the formation tester responses agree well
indicating that the Rotliegend dune/interdune sands interval (xl38-x232ft) contained
the best quality reservoir rock (porous and permeable)[Figure 4.1]. Above this interval,
the core and chippings show that siliceous, dolomitic and anhydride cements effect the
reservoir quality.

4-6
 Chapter 4: Analysis o f the Data Sets

The formation exposure time for the LWD tools (105 ft behind bit) were VA days
for the vertical well and 18 hours for the horizontal well. The LWD tools were assumed
to be working in the same manner for both the vertical and horizontal wells. However,
the LWD tools were seriously damaged during the drilling of the horizontal well, which
may have had a detrimental effect on the tool responses. The damage suggests that the
tools were not only being worn by abrasion, but also from being continuously knocked
against the borehole walls. For example, the memory port hatch was tom off during one
run. The variability in the LWD caliper (0.25 inches) proves some indication of the
amount of movement of the tools within the borehole. Also, there was a periodic
signature with a wavelength of 8ft (2.4m) throughout the horizontal well data.

4-7
 Chapter 4: Analysis o f the Data Sets

4.1.2 Data set 2

The data comprises log data sets from five wells; two vertical wells (wells 1, 33°
from vertical and 5, 21°), one horizontal sidetrack and two horizontal wells (wells lz, 3
and 6, all 90°). Core data were available for both vertical wells and one horizontal well
(well lz). The data used in this chapter from the five wells are summarised below in
Table 4.3. All the wells were drilled with OBM. Through the reservoir section, the
vertical wells/vertical sections (<75°) were drilled with 12lA inch bits. The horizontal
sections/wells were drilled with 8%inch bits. All three horizontal wells analysed in this
chapter were first drilled at medium angle (30-75°) through the reservoir. The well
trajectories were then turned to 90° through the B2 sand unit and drilled horizontally
through the reservoir, but advance up through the stratigraphy due to 5° bed dips.
Vertical wells 1 and 5 penetrate 221m (725ft) and 237.5m (779ft) of the
reservoir respectively. The medium angle sections of the horizontal wells 3 and 6
penetrate 139m (456ft) and 156.5m (513.5ft) of the reservoir respectively prior to
penetrating in the B2 unit [Table 4.4]. The horizontal reservoir sections of wells lz, 3
and 6 are 2076m (6811ft), 920m (3018ft) and 1528.5m (5015ft) respectively.
The gas and oil-bearing reservoir rocks are shale capped and the trap is a four­
way dip structure enclosed above a salt diapir. The gas leg is approximately 40m
(132ft) thick and the oil leg 58m (190ft). The horizontal wells were positioned
approximately two thirds into the oil leg below the GOC (gas/oil contact). Well 5 is a
gas injector in the centre of the field, the other four wells are oil producers [Figure 4.3].
The reservoir has been subdivided into three distinct units named A, B and C;
the unit tops are given below [Table 4.4]. The oil leg is contained within the B
sandstone, almost entirely within the B2 unit. The unit A3 is predominantly shaley and
may form a field wide permeability barrier reducing the effectiveness of the gas
injection. Unit A being uppermost, relates to the waning of the fan. Unit B represents
the most sand rich and thus productive unit consisting of amalgamated channel sands.
The lowermost unit C represents the early development of the fan, although
predominantly sand rich the unit is below the OWC. The sub-units relate to correlatable
features within the log data related to lobe switching within the fan. The GOC is
typically found within unit B1 and the OWC in unit B2. Geological description of the
field for data set 2 is given appendix 2.

4-8
Figure 4.3 Field map with well locations (provided by the Oil Company).
Top Palaeocene Reservoir Well Locations

□ Gas cap
Initial GOC = '496 m T V D SS

Oil leg
Initial OWC = 554 mTVDSS
Aquifer

Fault

Appraisal well

Well trajectory

Chapter 4: Analysis of the Data Sets

Horizontal oil producer

Gas injector
OWC monitoring
Proposed
 Chapter 4: Analysis o f the Data Sets

W eill

Well 1was drilled with OBM, both wireline and core data sets were available for
this vertical well [Figure 4.4 and Table 4.3]. The borehole was logged 9V2hours after
mud circulation stopped. The caliper values indicated that the borehole was smooth, but
slightly enlarged by approximately 0.2inches through the reservoir. The density
correction log was positive due to mudcake [Table 4.5]. Barite was used in the mud
(144ppb) resulting in the slightly high PEF values (2.5B/e) in the sandstone layers. The
gamma ray (>45API) and photoelectric factor (>3B/e) indicated that shale layers were
present through the reservoir. Low GR (<50API) and high PE (>4B/e) spikes were due
to doggers (thin limestone bed or sandstone with limestone cement, usually
significantly reducing porosity and permeability).
At the top of unit A1 (x571m)[Table 4.4], gas crossover was evident and the
induction resistivity values indicated the presence of hydrocarbons [Figure 4.4]. The Oil
company picked the GOC contact at x619m in unit A3. Above x652m in unit B1 the oil
was at irreducible water saturation marked by the irreducible oil contact (IOC) [Figure
4.4], The OWC contact was picked from the resistivity logs at x702m. There was little
separation of resistivity values through the reservoir showing only shallow mud
invasion.
The acoustic travel time and compressional acoustic travel time logs separate
(i.e. noisy) between x575.5-580m and x604-620.5m in shaley beds within the gas leg.
The acoustic travel time log values increase (> lOOps/ft) in comparison with the
compressional acoustic travel time log values (87-97ps/ft). The most plausible case
being weak beds, barring the possibility of poor data processing, because there was no
significant change in the neutron porosity or bulk density expected due to the shale
beds. Another feature of the both acoustic travel time logs was the low travel time
spikes at x659m and x679m. Both spikes were associated with low neutron porosity
values and high bulk density values. Comparison with the core records show that both
peaks resulted from doggers.

4-10
 Chapter 4: Analysis o f the Data Sets

D eviation (deg)
HUHmsm 33 40-65 90 21.5
m55-70 89.5

00

00
Sn
B it Size (inches) 12Vi” 12 Vi” 8 V2” 12 Vi” 12 Vi”

M ud Type OBM OBM OBM OBM OBM OBM OBM

Mud W eight (sg) 1.29 1.16 1.16 1.16 1.41 1.16 1.16

Cl concentration (mg/1) 150000 158000 192000 192000 141000 147000 147000

Mud oil water ratio 61:39 69:31 58:42 58:42 62:38 66:34 66:34

Top depth o f data (m ) x5 0 0 x300 x500 x721 x500 x652 x957

B ottom depth o f data (m ) x7 8 9 x2376 x721 x l6 3 6 x757 x960 x2348

Gamma Ray WL LW D + WL LW D W L + LW D WL LW D WL

Caliper WL WL WL WL

D eep induction resistivity WL WL WL WL

Medium induction resistivity WL WL WL WL

Bulk Density WL LW D LW D WL + LW D WL LW D

D ensity Correction WL LW D LWD W L + LW D WL LW D

Photoelectric Factor WL LW D LW D. WL + LW D WL

Neutron Porosity WL LW D LW D. W L + LW D WL LW D

Acoustic Travel T im e WL WL WL

Compressional Travel Tim e WL WL

Shear Travel Time WL

Stoneley Travel Time WL

Electrom agnetic R esistivity LW D LW D LW D LW D

Tension WL WL WL. WL

Pressure LW D

Rate o f Penetration LW D LWD LW D LW D

Core Porosity Core Core Core

11
Core Horizontal Permeability Core Core Core

Core Vertical Permeability Core Core Core

Core Water Saturation Core Core Core

Table 4.3 Summary of well log data used in this chapter (WL = wireline).

BS38P3
A1 (sst w . m st) X571.00 X529.00 X1588.00 x 5 19.50 X652.00 x2337.00

A2.1 (clean sst) X581.21 X552.00 x 1529.00 X538.50 X667.00 X2301.00

A 2.2 (sst w . m st) X604.67 x 9 16.00 X592.00 xl473.00 X550.00 X746.00 x 2 186.00

A 3 (sst w . m st) x 6 16.87 x l 104.50 X627.00 x 1407.00 X568.00 X773.00 X1745.00

B 1 (clean sst) X630.02 xl333.00 x638.00 x 1283.00 x587.00 X790.50 X1691.50

B 2 (sst w. m st) X653.96 xl851.50 X668.00 x 1069.00 X593.50 X808.50 x 1639.00

C (sst w . m st) X760.00 X685.00

Table 4.4 Oil company formation tops (sst = sandstone, mst = mudstone).

4-11
 Chapter 4: Analysis o f the Data Sets

R H O B ( ^ c r i? ^ ^ ^ 2.286 0.094 2.112 2.703

RHOB2 (g/cm3) 2.296 0.083 2.164 2.706
NPHI (n rW ) 0.182 0.052 0.050 0.363
ILM (Ohm.m) 12.714 20.715 0.411 98.669
ILD (Ohm.m) 13.465 23.379 0.360 105.151
Rt (Ohm.m) 13.454 23.394 0.360 104.956
GR (API) 48.529 19.467 29.837 199.963
PEF (B/e) 3.025 0.679 2.062 6.002
DT (us/ft) 92.611 12.565 61.293 224.257

DTCO (ps/ft) 90.338 5.759 61.638 113.316

DRHO (g/cm3) 0.025 0.013 -0.037 0.082
CALI (inches) 12.519 0.191 12.420 13.346

CPOR (m3/m3) 0.205 0.045 0.024 0.266

Kv (mD) 226.171 190.762 0.010 810.000
Kh (mD) 244.553 219.548 0.010 1320.000
Grain Density (g/cm3) 2.647 0.038 2.560 3.150
Sw (%) 14.913 17.013 0.700 94.400

Table 4.5 Well 1 wireline descriptive statistics.

Well lz

Well lz was drilled with OBM, LWD and core data sets were available for this
horizontal well [Table 4.3]. The LWD data acquired was in the 8 V2” hole section and
was logged approximately 2.3hours behind the bit (measured from the LWD
density/neutron tool measurement point). The rate of penetration was approximately
15m/hr throughout drilling the well [Table 4.6]. The logs were cyclical due to borehole
spiralling on a wavelength of l- 2 m.
The LWD density correction was close to zero, yet highly erratic with some
significant spikes in the data (>+0.1g/cm )[Figure 4.5]. Notable spikes occur between
xll85-1240m (due to caving), x2023-2072m and x2209-2217m, both due to drilling
difficulties, as the formation becomes muddier. The drilling fluid report noted that the
mud weight was increased early in drilling the well. It was assumed for the purposes of

4-12
 Chapter 4: Analysis o f the Data Sets

this thesis that barite was used in the mud considering that 144ppb barite was used in
the drilling the vertical well 1 .
The Oil Company considered the LWD bulk density values to be of poor
quality. The LWD bulk density values were much lower than expected from the vertical
well 1 [Figure 4.5]. The LWD bulk density log appears to be greatly effected by
variable standoff of the tool. The GOC gradational contact was picked between
xl547.0-1596.0m (Unit Bl).
The neutron porosity values were greater than expected from well 1. This was
likely due to the standoff, but invasion could also be a factor. The drilling report noted
that large quantities of mud were lost during the drilling and therefore invasion was
likely to be deep. However, the match between core and neutron porosity was good,
contradicting standoff as the cause of the ‘high’ neutron porosity values [Figure 4.5].
The LWD resistivity values were slightly lower than the resistivity values
recorded in the vertical well 1 [Figure 4.5]. This was likely to be due to differences in
the operating frequencies of the tools used. The peaks observed in the pressure log
(DDS) were generally due to bit trips (removing the entire drill pipe to change the bit
prior to continuation of drilling). The rate of penetration was low throughout most of
the well, large increases were noted in shalier sections of the borehole.

Well 3

Well 3 was drilled with OBM and logged with wireline (pipe conveyed) and
LWD through the Scinch horizontal well. The rate of penetration was between 10-
20m/hr [Figure 4.6]. The LWD amplitude resistivity log “straight lines” [e.g. x640-
655m], indicating that the measurement conditions were exceeded, the phase resistivity
log was not similarly effected. The LWD nuclear logs were noisy due to a rugose well
and barite in the mud. The gradational GOC was picked between x647.0-652.5m (Unit
Bl).
From the wireline logs, the caliper values show that the borehole was rugose
with several caves and ledges and all the logs were cyclical due to l- 2 m borehole
spiralling, especially between x850-1010m resulting from a damaged bit. The wireline
density correction log was close to zero, but negative excursions occurred in muddier
beds and doggers. The drilling fluid report noted that barite mud was used, resulting in
high and erratic PEF values (2.3-2.5B/e) in the sandstone layers with large departures

4-13
 Chapter 4: Analysis o f the Data Sets

due to the shale beds and doggers. Bulk density and neutron porosity values were
comparable in values with the well 1 values [Figure 4.4, Figure 4.6b, Table 4.5 and
Table 4.7].

2.104 0.113 1.650 3.110

NPHI (m3/m3) 0.233 0.010 0.214 0.329
EWR (Ohm.m) 9.543 15.013 2.570 650.270
Rt (Ohm.m) 9.543 15.013 2.570 650.270
GR1 (API) 15.307 2.765 9.440 34.300
GR2 (API) 28.557 9.194 13.740 88.390
DRHO (g/cm3) -0.002 0.040 -0.430 0.330
ROP (m/s) 0.025 0.058 0.000 0.939
DDS (psi) 12392.045 40460.029 684.000 365939.969

CPOR (m3/m3) 0.240 0.011 0.215 0.289

Kv (mD) 237.311 78.398 4.930 416.000
Kh (mD) 229.204 67.583 24.500 345.000
Grain Density (g/cm3) 2.640 0.007 2.590 2.650
Sw (%) 11.285 3.526 2.800 18.100

Table 4.6 Well lz LWD descriptive statistics.

R H O B ^cm 3^ ^ 2.299 0.105 2.056 2.870

NPHI (m3/m3) 0.208 0.052 0.001 0.635
ILD (Ohm.m) 18.722 37.569 1.506 268.564
ILM (Ohm.m) 15.756 23.518 1.255 445.156
Rt (Ohm.m) 15.594 23.280 0.934 364.724
SGR (API) 41.781 17.793 14.258 116.542
PEF (B/e) 2.596 0.773 1.821 9.977
DRHO (g/cm3) -0.001 0.029 -0.167 0.343
CALI (ins) 8.484 0.244 7.601 10.744
ROPS (m/hr) 22.390 6.128 6.293 45.660

Table 4.7 Well 3 wireline descriptive statistics.

4-14
 Chapter 4: Analysis o f the Data Sets

Well 5

Well 5 was drilled with OBM, both wireline and core data sets were available
for this \2 lA inch vertical well [Figure 4.7 and Table 4.3]. The borehole was logged
19hours after mud circulation stopped, except the acoustic and dual-arm caliper tools
that were logged after 44V2 hours.
The density caliper values indicated that the borehole was rugose, but in-gauge,
however the dual-arm caliper demonstrated that the borehole was slightly enlarged by
approximately 0.1 inches along the major axis followed by the density caliper [Figure
4.7]. The other arm of the caliper (perpendicular to the first) revealed that the minor
axis of the borehole was less rugose and approximately linch under-gauge, an
indication of the unconsolidated nature of the sandstone, although mudcake could not
be ruled out as the cause. The density correction log was slightly negative due to barite
(1961b/bbl) and the rugose borehole [Figure 4.7]. The effect of the barite was not as
great on the PEF log values as expected from well 1 [Figure 4.7 and Figure 4.4].
The gradational GOC was picked between x597.0-600.0m (Unit B2). For the
purposes of this thesis the OWC was picked at x665.5m. The oil column was at
irreducible water saturation (IOC) above x640m. There was some separation of
resistivity values through the reservoir, due to mud invasion (>10inches)[Figure 4.7].
The acoustic travel time log values were near constant (>90ps/ft) through the units A2.2
to B2. The shale beds in units A1 and A2.1 appeared to be ‘fast’ (competent), but the
sandstone beds were ‘slow’ probably due to gas, but the unconsolidated beds may be a
factor. There were no doggers readily apparent through the reservoir.

4-15
 Chapter 4: Analysis o f the Data Sets

Well 6

Well 6 was drilled with OBM, the datasets included wireline (pipe conveyed)
and LWD for this Scinch horizontal well. However, there were large gaps in the LWD
data set [Table 4.3]. The LWD bulk density and neutron porosity data only covered
x770-815m, which allowed the gradational GOC to be picked between x797-804m (Oil
Company depths) [Figure 4.8b]. Comparison of the resistivity logs, suggested that well 6
remains in the oil leg over the interval covered by wireline data [Figure 4.8a].
There was no wireline caliper log for this well, however, the tension provided
some indication of the downhole conditions whilst logging. The tension log was erratic
in comparison with the tension log of well 5 [Figure 4.7 and Figure 4.8a]. The erratic
nature was a result of the nature of pipe conveyed logging being a stop/start operation
and the rugose borehole condition, though could result from poor drilling crew
operation of the pipe conveyed logging (PCL). Often PCL depths are poor due to
inability to keep the wireline taut at all times directly effecting the quality of the depth
derived logs such as acoustic and induction. Additionally, the Stoneley wave travel time
could be used as an indicator of poor hole conditions. Large increases in Stoneley wave
travel time may indicate the presence of fractures or incompetent beds [x 1955-1990m
Figure 4.8a].
The wireline gamma ray and resistivity values were similar to the expected
values from the other wells [Figure 4.8a]. However, the compressional travel time log
was approximately lOps/ft slower through the sandstone beds compared with wells 1
and 5 [Figure 4.8a, Figure 4.4 and Figure 4.7]. The shale beds were faster than the
sandstone beds and effected the compressional travel time log to a greater extent than in
both wells 1 and 5. This was possibly a result of shale acoustic anisotropy or may be
evidence of sand acoustic anisotropy. Another possibility was that the acoustic tool was
not running parallel with the borehole. This can produce both faster and slower travel
times than expected. Well 6 also encountered many more doggers than observed in the
other wells [Figure 4.8a]. However, some of the fast peaks correlated with the tension
log and suggested that those peaks were a result of noise induced by the pipe conveyed
logging operation [xl060m Figure 4.8a].

4-16
 Chapter 4: Analysis o f the Data Sets

R H O B (^ c r r ? ^ ^ ^ ^ 2.279 0.124 2.038 2.825

NPHI (pu) 0.196 0.062 0.027 0.401
AT10 (Ohm.m) 16.596 29.193 0.347 170.347
AT20 (Ohm.m) 16.004 28.549 0.320 161.892
AT30 (Ohm.m) 15.570 27.788 0.323 156.529
AT60 (Ohm.m) 15.505 28.147 0.339 215.159
AT90 (Ohm.m) 15.120 27.028 0.342 218.142
Rt (Ohm.m) 15.120 27.036 0.342 218.142
GR (API) 57.095 29.537 22.705 218.019
PEF (B/e) 1.995 0.473 0.953 4.213
DT (|L is/ft) 90.501 8.845 49.199 129.771
DRHO (g/cm3) -0.006 0.016 -0.048 0.074
CAL1_1 (ins) 11.231 0.123 11.179 11.647
CAL2_1 (ins) 12.308 0.124 12.127 13.143
CALI_DEN (ins) 12.348 0.354 12.105 14.215
TENS (lb/ft) 4052.401 104.119 3856.000 4443.000

CPOR (m3/m3) 0.238 0.026 0.181 0.275

Kh (mD) 257.933 159.429 2.800 536.000
Grain Density (g/cm3) 2.648 0.004 2.640 2.650

Sw(%) 37.472 21.219 12.700 83.500

Table 4.9 Well 5 wireline descriptive statistics.

IDPH (Ohm.m) 15.689 11.440 0.256 558.333

ILD (Ohm.m) 15.018 8.170 0.476 450.221
ILM (Ohm.m) 14.026 6.991 1.039 281.647
IMPH (Ohm.m) 14.429 7.774 1.051 288.121
GR (API) 35.073 14.168 15.171 102.990
DTCO (ps/ft) 92.328 10.745 50.236 122.951

DTSM (ps/ft) 156.311 11.921 100.006 222.767

DTST (ps/ft) 284.992 7.204 257.078 352.263

TENS (N) 11066.921 1726.296 3131.548 21137.949

Table 4.10 Well 6 wireline descriptive statistics.

4-17
 Chapter 4: Analysis o f the Data Sets

The wallet at the back of the thesis contains all the log plots.

Figure 4.4 Well 1 wireline logs.

Figure 4.5 Well lz LWD logs.

a) Well 3 Wireline logs

b) Well 3 LWD logs

Figure 4.7 Well 3 logs.

Figure 4.8 Well 5 wireline logs.

a) Well 6 Wireline logs

b) Well 6 LWD logs

Figure 4.10 Well 6 logs.

4-18
 Chapter 4: Analysis o f the Data Sets

4.2 Porosity Anomalies

4.2.1 Porosity anomalies - data Set 1

Calculating porosity for clean sandstone reservoirs in a vertical well is usually

relatively straightforward, even for gas reservoirs. Once the correct hydrocarbon density
has been estimated/calculated, the hydrocarbon corrections (usually only fluid density is
needed) can be calculated and porosity determined. Core porosity values (if available)
are used to calibrate the porosity values from the logs (§2.2). This approach provides
accurate and reliable porosity values from the logs.
In horizontal wells, the problem of calculating porosity is less straightforward.
Often the borehole and formation conditions are significantly different from the
situation found in vertical wells (See Chapter 3) leading to incorrect porosity estimation
when the vertical well approach is used. Incorrect porosity estimation is caused by
perturbations to the log measurement, most likely as result of one or all of the
following: anisotropic mud invasion, borehole damage and tool orientation.
For the data set 1, the porosity estimation in the vertical well was simple.
However, porosity estimation was more complicated in the horizontal well. In the
section above, borehole and tool damages were mentioned and the former is self-
evident from the logs, especially the caliper logs that show cyclic rugosity [Figure 4.2].
Permeability anisotropy [Table 4.10, Figure 4.9 and Figure 4.10] and therefore mud
invasion anisotropy is well known for the Leman sands (Bedford et al. 1997). Most
reported cases of unusual porosity values in horizontal wells discuss estimated porosity
that is ‘too high’ [Equation 2.2]. In this case study, unusually the porosity values
calculated were too low and this was of concern to Oil Company 1.

Horizontal Vertical Overburden Corrected Grain Density

Permeability (mD) Permeability (mD) Core Porosity (%) (g/cm3)
Mean 5.18 1.86 11.41 2.66
SD 13.38 5.31 3.32 0.02
Min 0.02 0.01 5.40 2.63
Max 88.00 40.00 19.40 2.79

Table 4.10 Data set 1 vertical well core data descriptive statistics.

4-19
 Chapter 4: Analysis o f the Data Sets

Permeability Histogram
50

— K h (m D )

— K v (m D )

0.01 0.1 10 100

Perm eability (mD)

Figure 4.9 Vertical well core permeability histogram. Note: the horizontal and vertical
permeability distributions are very similar, but when plotted against porosity [Figure
4.10], permeability anisotropy is evident.

Porosity verses Log Permeability

2.5
* Log Kv (m D ) • LogKh(mD)

— Linear (Log Kv (m D )) — Linear (Log Kh (m D ))

1.5 •
■-
■ *

S 0.5 ♦ ^ ♦

y = 0.2283x- 2.7478
♦
I R2 = 0.8534
|- 0 .5 ■ i :j r - - ••
t ♦ V V *
4* • y = 0 .1 8 7 8 x -2.565
t
* ■
♦
* 44 ♦
♦ R2 = 0.8217
-1.5
♦4 ♦ ♦

-2.5
10 15 20
Porosity (pu)

Figure 4.10 Vertical well core porosity permeability plot.

4-20
 Chapter 4: Analysis o f the Data Sets

Crossplotting horizontal well LWD standard model porosity values [Table 4.11]
against depth distributed core porosity values (Weissliegend and Rotliegend gas only)
from the vertical well show that the match was poor especially with LWD average
density porosity [Figure 4.11c]. However, there was substantial scope for altering the
position of the core porosity values in depth with respect to the horizontal well porosity
data and thereby changing the match achieved. After testing a number of different
approaches, the best comprise was to evenly distribute the core porosity values in depth.
Despite this potential for compromise of the comparison the horizontal and
vertical well density porosity data still clearly demonstrate low horizontal well density
porosity values compared with the vertical well [Table 4.12]. This is clear even though
the rugosity, indicated by the greater standard deviation, acts to increase the porosity
values in the horizontal well.
Comparison of the neutron porosity mean values would suggest that the
horizontal and vertical wells were in good agreement [Table 4.12], but as for the density
porosity values, rugosity acts to increase the recorded porosity values. Section 4.4
shows that filtering to reduce the effect of rugosity lowers both the mean and standard
deviation of the horizontal well neutron porosity values.
The observations of low porosity values presented form the basis of the
investigation of the horizontal well porosity values for data set 1. The calculation of
reserves is very sensitive to the porosity value and low values may suggest that the field
could be uneconomic to develop or to continue development.

Porosity Parameter Standard values

Fluid Density (g/cm3) used for density logs 1.000
Excavation correction (percentage porosity) used for neutron porosity logs 0.000

Table 4.11 Parameters used for porosity calculations.

4-21
 Chapter 4: Analysis o f the Data Sets

V ariable M inim um Maximum M ean S tandard Deviation

A verage Density Porosity (% ) -0.005 0.226 0.137 0.035

M axim um Density Porosity (% ) -0.020 0.225 0.131 0.036
N eutron Porosity (% ) 0.020 0.157 0.095 0.031

H orizontal Helium Porosity (% ) 0.054 0.194 0.114 0.033

Acoustic Porosity (% ) 0.028 0.213 0.106 0.044

Density Porosity (% ) 0.011 0.201 0.112 0.035
N eutron Porosity (% ) 0.013 0.144 0.081 0.034

Average Density Porosity (% ) -0.092 0.450 0.193 0.073

M aximum Density Porosity (% ) -0.249 0.275 0.085 0.078
Neutron Porosity (% ) -0.034 0.301 0.083 0.033

Table 4.12 Derived porosity statistics (fluid: lg/cm3 density and 189ps/ft travel time).

Vertical Well Wireline Porosity Crossplot

0.25
Acoustic Porosity

Density Porosity

Neutron Porosity
0.2 -* * ■

*x

0.15

Xx

x<
0.1

0.05

0
0 0.05 0.1 0.15 0.2 0.25
Core Porosity (%)

a)

4-22
 Chapter 4: Analysis o f the Data Sets

Vertical Well LWD Porosity Crossplot

0.25
LWD Average Density
Porosity

LWD Maximum Density

0.2 -
Porosity

LWD Neutron Porosity

xx

0.05

0 0.05 0.1 0.15 0.2 0.25

Core Porosity (%)

b)

Horizontal Well LWD Porosity Crossplot

0.45 X

0.4
X
0.35
X

g 03
XXX
X
X

* X
|S0-25 X
o x *
0. 0.2 X % X> /
*x X
cxxxx
'0.15 /
X x L W D A v e ra g e D e n s ity
X^<x; P o rosity
0.1
<
X<X x L W D M axim um D ensity
/x p X P o ro sity
0.05
X L W D N eutron P o ro s ity
X
0 / l I i
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45
Core Porosity (%)

c)

Figure 4.11 Core porosity against calculated porosity crossplots using standard
parameters for vertical well a) wireline and b) LWD data, c) horizontal well LWD data.

4-23
 Chapter 4: Analysis o f the Data Sets

4.2.2 Porosity anomalies - data Set 2

Data set 2 required a different approach to data set 1 for calculating porosity in
the mixed-fluid shaley sandstone reservoir. The shale needed to be accounted for and
the correct hydrocarbon densities had to be estimated/calculated. The shale parameters
used for data set 2 are given in Table 4.14. Core porosity values were used to calibrate
the porosity values from the logs (See Chapter 2).
In the horizontal wells, calculating porosity was problematic due to borehole
and formation conditions differing from the vertical wells (See Chapter 3). A further
complication was that horizontal wells are not logged with the same class of tools
(wireline/LWD) or the same suite of tools (different measurements)[Table 4.3].

Shale Parameter Values

Shale density (g/cm3) used for density logs 2.500

Shale point (percentage porosity) used for neutron porosity logs N ot applicable

Shale travel time (ps/ft) used for acoustic travel time logs 100.000

Table 4.14 Parameters used for shale corrections.

A number of porosity algorithms were used: shale corrected density, shale

corrected acoustic [Equations 2.32 and 2.34] and sandstone neutron porosity. Shale
corrected neutron porosity was not used due to over correction of the ‘shale’ effect in
the sands. Log and core porosity values were compared in vertical and horizontal wells
to determine if the log porosity over estimated porosity in horizontal wells with
parameters used given in Table 4.15 and the calculated porosity values in Table 4.16.
A direct comparison of porosity values between wells was not necessarily valid
due to lateral variations in reservoir characteristics and cross-section of the reservoir
intersected by the wells differs. There was little reason to assume that porosity was
laterally constant, since the sands were shaley and turbiditic with several doggers.
However, without additional core porosity data no conclusive assessment of porosity
homogeneity could be made. For example, well 1 core porosity was sampled every
25cm through the reservoir (-450 samples), but <20 samples spread through the
reservoir were available for well 5. More than 80 samples from a single bed were
available for horizontal sidetrack well lz. A decision was made to use well 1 core

4-24
 Chapter 4: Analysis o f the Data Sets

porosity data for the statistical comparisons in section 4.4 in all but well lz where the
large number of core samples permitted their use.

Porosity P aram eter S tandard values

Fluid density (g/cm3) used for density logs 1.000
Excavation correction (percentage porosity) used for neutron porosity logs 0.000
Fluid travel tim e (ps/ft) used for acoustic travel time logs 189.000

Table 4.15 Parameters used for porosity calculations.

RHOB 0.221 0.057 0.326 -0.032

RHOB2 0.215 0.050 0.295 -0.034
NPHI 0.182 0.052 0.050 0.363
DT 0.281 0.094 0.047 1.263
DTCO 0.264 0.043 0.050 0.435
RHOB 0.216 0.042 0.305 0.027
RHOB2 0.219 0.040 0.295 0.026
NPHI 0.180 0.029 0.050 0.272
DT 0.252 0.034 0.047 0.350
DTCO 0.241 0.036 0.050 0.321
CPOR 0.205 0.045 0.024 0.266
CPOR 0.198 0.055 0.054 0.266

RHOB 0.331 0.068 0.606 -0.279

NPHI 0.233 0.010 0.214 0.329
RHOB 0.375 0.064 0.606 -0.279
NPHI 0.240 0.009 0.225 0.329
CPOR 0.240 0.011 0.215 0.289
CPOR 0.239 0.010 0.215 0.289

RHOB 0.213 0.064 0.360 -0.133

NPHI 0.208 0.052 0.001 0.635
RHOB 0.213 0.059 0.293 -0.061
NPHI 0.209 0.043 0.001 0.394

RHOB 0.225 0.075 0.371 -0.106

NPHI 0.196 0.062 0.027 0.401
DT 0.265 0.066 -0.043 0.558
RHOB 0.256 0.041 0.360 0.060
NPHI 0.198 0.029 0.047 0.283
DT 0.263 0.024 0.199 0.341
CPOR 0.238 0.026 0.181 0.275
CPOR 0.238 0.026 0.181 0.275

DTCO 0.279 0.080 -0.036 0.507

DTCO 0.283 0.099 -0.031 0.501

Table 4.16 Data set 2 comparison of porosity descriptive statistics in m3/m3. Values
from the whole reservoir are in white cells. Values from the unit B2 are in grey cells.

4-25
 Chapter 4: Analysis o f the Data Sets

Comparison of the porosity values in Table 4.16 shows that for the horizontal
wells lz, 3 and 6 all measure greater porosity values (>1 standard deviation) in the
whole reservoir section and unit B2 (least shaley unit) when compared with the vertical
wells 1 and 5 log and core porosity values. Notably the acoustic porosity values were
significantly greater than other porosity measurements.
The high observed porosity values could result in over optimistic reserves
estimates from the horizontal well data. Oil Company 2 was developing the field
predominantly with horizontal wells therefore inaccurate porosity was of concern. This
evidence of high porosity values in the horizontal wells of data set 2 forms the basis of
the investigation presented below in section 4.4.

4-26
 Chapter 4: Analysis o f the Data Sets

4.3 Analysis techniques

This section provides an explanation of how the analyses of the horizontal well
porosity values were performed. The first section (§4.3.1) describes the approaches
used for the porosity calculations and the decision process used to choose the best
method is discussed in section 4.3.2 with a worked example from each data set. Finally,
a description of the spectral analysis of the data used is provided with a worked
example from each data set, used to reduce the effects of borehole conditions resulting
from spiralling. The results of the analyses of the two data sets are presented in the
following section 4.4.

4.3.1 Porosity algorithms

The porosity algorithms used for the analysis were detailed in chapter 2. The
porosity algorithms used were density [Equation 2.2], neutron [Equation 2.11], acoustic
[Equation 2.15], square root [Equation 2.46], Wiley & Patchett [Equation 2.44] and
hydrocarbon corrected [Equations 2.50 and 2.51] using iterative calculations in which
the fluid parameters were used to achieve good agreement with core porosity values as
defined in section 4.3.2. Section 2.4.3 described the Gaymard and Poupon (1968),
Wiley and Patchett (1994) and Cowan and Wright (1997) methods. Porosity histograms
and statistical comparisons with core porosity values were used to determine the
validity log porosity values and test which method provided the most reliable values as
detailed in section 4.3.2. A different approach for porosity calculation was used for each
data set due to the different lithologies and fluid contents of the reservoirs.

Data set 1

The field from which data set 1 was obtained consisted of the Leman gas sand
of the southern North Sea. For the purposes of this thesis the reservoir was subdivided
into three units based on the vertical well; Weissliegend x008ft-xl53ft, Rotliegend gas
xl53ft-xl91ft and Rotliegend water xl91ft-x236ft [Figure 4.1, layers marked]. This
was based on the change in facies from the Weissliegend to the Rotliegend (at x 153ft)
and the change in formation fluid from gas to water (at xl91ft). The porosity analyses

4-27
 Chapter 4: Analysis o f the Data Sets

were performed on the whole reservoir and the three units separately for each of the
porosity algorithms in turn. The porosity algorithms used were density, neutron,
acoustic, square root, Wiley & Patchett and iterative - hydrocarbon corrected
calculations. Note that the Rotliegend water unit was not observed in the horizontal well
so only the other two units were used in this case.
In the vertical well, it was possible to calculate porosity values from the density,
neutron and acoustic logs from both the wireline and LWD (no acoustic) tool
measurements using the methods described above. The objective was to determine
which method/s were sufficient to calculate accurate porosity values, since there is no
point in practice opting for a complex method over a simple method without any
notable improvement in accuracy. It was also important to find satisfactory LWD
derived porosity to core porosity agreement because no wireline logs for the horizontal
well exist.
Initially, the density derived porosity values (both wireline and LWD) were
calculated assuming a matrix density = 2.65g/cm (used for all density porosity
calculations) and fluid density = lg/cm . The neutron derived porosity values (both
wireline and LWD) were calculated in sandstone units assuming no corrections were
required. The acoustic derived porosity values (only wireline) were calculated assuming
the matrix acoustic slowness = 55psec/ft (used for all acoustic porosity calculations)
and fluid acoustic slowness = 189psec/ft. This approach was applied to the whole
reservoir section and was called the standard model in this thesis for data set 1.
The single layer method used to improve the calculation of log porosity was a
simple linear shift over the whole interval of interest i.e. finding the optimum fluid
density, neutron porosity excavation constant (a constant added to account for gas) and
acoustic fluid travel time. The core porosity data was used to minimise the difference
between the core porosity values and the appropriate log derived porosity values by
changing the appropriate fluid parameter (using SOLVER™ for EXCEL™). The three-
layer model used different fluid density, neutron porosity excavation constant and
acoustic fluid travel time (by minimising the core and log derived porosity differences
by changing the appropriate fluid parameter) for each of the three layers defined above.
The square root and Wiley & Patchett methods were applied to the neutron and density
porosity values calculated from the standard, single and three-layer models above for

4-28
 Chapter 4: Analysis o f the Data Sets

both LWD and wireline porosity values. The hydrocarbon corrected method used the
square root and Wiley & Patchett equations as detailed in section 2.4.3.
In the horizontal well, porosity values were calculated from the LWD density
and neutron tool measurements. The same methods for calculating porosity were used
as for the vertical well, but without the benefit of horizontal well core porosity data to
verify the LWD calculated porosity values. In addition, only two of the three units were
observed in the horizontal well; Weissliegend xl34-906ft and xl655-1951ft, and
Rotliegend gas x906-1655ft [Figure 4.2, layers marked].

Data set 2

Data set 2 required a slightly different approach than was used for data set 1
because of the shaley mixed-fluid sandstone reservoir. The shale needed to be
accounted for [Equation 2.33] and the correct hydrocarbon densities had to be
estimated/calculated (See Chapter 2). Core porosity values were used to calibrate the
porosity values from the logs (See Chapter 2). For the purposes of this thesis, the
reservoir was subdivided into four units based on the change in formation fluid from
gas to oil to water [Table 4.17]. The porosity analyses were performed on the whole
reservoir and the four units separately for each of the porosity algorithms in turn. A
number of porosity algorithms were used; shale corrected density, shale corrected
acoustic, square root, Wiley & Patchett, Gaymard & Poupon neutron and Gaymard &
Poupon density [Equations 2.37, 2.39, 2.46, 2.52, 2.50 and 2.51 respectively] and
sandstone neutron porosity. Shale corrected neutron porosity was not used due to over
correction of the ‘shale’ effect in the sands.
The fluid parameters for the density and acoustic shale corrected porosities were
investigated in two ways. Firstly, using the Oil Company’s (OC) fluid parameter values
and secondly with values such that the porosity equations are in agreement with well 1
core porosity values by minimising the core and log derived porosity differences by
changing the appropriate fluid parameter. Porosity values were calculated from the logs
available from the well; wireline and/or LWD measurements of density, neutron or
acoustic logs using the methods described above. The logs were used to determine
which method/s were sufficient to calculate accurate porosity values for both LWD and
wireline horizontal well data when compared with well 1 core porosity values.

4-29
 Chapter 4: Analysis o f the Data Sets

x 1547-2270
T ran sition

Table 4.17 Top depths of fluid zones for each well in metres.

The Oil Company’s porosity calculation method was reproduced. The density
derived porosity values (both wireline and LWD) were calculated using the matrix
density = 2.65g/cm3, shale density = 2.5g/cm3 and water density = lg/cm3 and
hydrocarbon density = 0.7g/cm3 [Equation 2.37]. The neutron derived porosity values
(both wireline and LWD) were calculated in sandstone units assuming no corrections
were required. The acoustic derived porosity values (only wireline) were calculated
assuming the matrix acoustic slowness = 55psec/ft and the shale acoustic slowness =
lOOpsec/ft (used for all acoustic porosity calculations) and fluid acoustic slowness =
189psec/ft [Equations 2.37]. The Oil company shale parameters were used for all
density and acoustic porosity calculations.
The method used to improve the calculation of log porosity was a simple linear
shift over the fluid unit of interest i.e. finding the optimum fluid density, neutron
porosity excavation constant (a constant added to account for gas) and acoustic fluid
travel time. The core porosity data was used to minimise the difference between the
core porosity values and the appropriate log derived porosity values by changing the
appropriate fluid parameter (using SOLVER™ for EXCEL™). The square root and
Wiley & Patchett and hydrocarbon corrected methods were applied to the neutron and
density porosity values calculated from the optimised shift values above for both LWD
and wireline porosity values. Porosity histograms and statistical comparisons with core
porosity values were used to determine the validity of the hypothesis that the calculated
porosity values were over estimated in horizontal wells (§4.2.2).

4-30
 Chapter 4: Analysis o f the Data Sets

4.3.2 Statistical tests

Porosity histograms and statistical comparisons with core porosity values were
used to determine the validity of the hypothesis that log porosity values were
anomalous in the horizontal wells of both data sets. Using data set 1 the horizontal well
was tested to verify that the porosity values were under estimated in comparison with
the vertical well core porosity values. Using data set 2 the horizontal wells were tested
to verify that the porosity values were over estimated in comparison with the vertical
(well 1) core porosity values (except horizontal well lz which was tested against core
porosity values from well lz itself). The vertical wells of both data sets were also
compared with core in the same manner to confirm the hypothesis that the horizontal
and not the vertical well log porosity values were anomalous for both data sets.
Porosity histograms allowed a direct visual comparison of the core and log
derived porosity values, whilst the statistical tests used, Student’s t and Chi2 tests,
enabled quantitative assessments of the comparisons (examples given below). Student’s
t-tests were used for all cases where F-tests demonstrated their distributions closely
approximated a normal distribution. F-tests of the standard deviations of the
distributions were used to verify the use of Student’s t-tests was robust; in all other
cases Chi2 tests were used. For Student’s t-tests, the hypothesised difference in the
means being equal to zero and unequal variance are assumed. For whole reservoir
cases, the Chi2 goodness of fit test was used due to the non-normal distributions evident
from the histogram plots of the data (see §4.4 and examples given below). The
statistical tests were all comparisons of the core porosity values against the log derived
porosity values as stated for each test in section 4.4 and the examples below.
For the purposes of this thesis, good agreement between porosity values of a
given model and the core porosity values is defined as when the result of an appropriate
two-tailed statistical test was within the 95% confidence limit. Tables in section 4.4 and
the examples below detail the Student’s t-tests for clear cells and Chi2 tests for the grey
cells. A tick or cross ( / IX ) represents acceptance or rejection of the hypothesis, at the
95% confidence level, that there is zero difference between that porosity equation and
core porosity values over the unit stated.
Correlation coefficients were also calculated to provide an indication of the
relative spatial relationship between core and calculated porosity values over the whole

4-31
 Chapter 4: Analysis o f the Data Sets

reservoir. Correlation was not valid between the vertical well core porosity and
horizontal well calculated porosity values because the data were not spatially related.
A direct comparison of porosity values between wells was not necessarily valid
(§4.2.1). For data set 1, crossplotting horizontal well LWD standard model porosity
values against depth distributed core porosity values (Weissliegend and Rotliegend gas
only) from the vertical well showed that the match was poor especially with LWD
average density porosity [Figure 4.11].
For data set 2, well 1 core porosity was sampled every 25cm through the
reservoir (~450 samples), but <20 samples spread through the reservoir were available
for well 5. For the horizontal sidetrack (well lz) there were >80 samples from a single
bed. A decision was made to use well 1 core porosity data for the statistical
comparisons in all but well lz where the large number of core samples permitted their
use (§4.2.2).
The lack of core porosity data was problematic for accurate determination of
which method(s) were appropriate to calculate accurate horizontal well porosity values.
However, comparison of log calculated porosity with the scaled vertical well core
porosity values over the relevant units allowed rejection of methods which were
certainly inaccurate.
A further complication in the horizontal wells was the borehole and formation
conditions differing from the vertical wells (See Chapter 3). Calculating porosity was
problematic because the horizontal wells were not logged with the same class of tools
(wireline/LWD) or the same suite of tools (different measurements)[Table 4.3]. This
was due to operational concerns rather than petrophysical needs and can frequently
result in poor quality horizontal well data.

4-32
 :'l
!;!
Chapter 4: Analysis o f the Data Sets j!
| ,
Data set 1 example |

The results of the standard porosity analyses are presented for the vertical well
[Figure 4.16, Table 4.19 and Table 4.20] and the fluid parameters used are listed in
Table 4.17. The porosity calculations in the vertical well clearly show that porosity
calculations were most reliable using wireline density values. The wireline density
porosity values were in good agreement with core porosity values over all depth
intervals and have the highest correlation with the core porosity values using the
standard model [Figure 4.16, Table 4.19 and Table 4.20]. The wireline neutron and
LWD maximum density porosity values were only in good agreement with core
porosity values over the Rotliegend water intervals. The wireline acoustic and LWD
neutron porosity values were in good agreement with core porosity values over the
whole and Rotliegend water intervals. LWD average density porosity values were in
poor agreement with core porosity values over all depth intervals.
The vertical well was logged with LWD (LAD) 2 days AFTER the wireline
tools. LWD neutron, average and maximum density porosity values were greater than
their respective wireline porosity values. The LWD neutron porosity values were
increased with respect to the wireline values due to the increased depth of mud
penetration increasing the hydrogen index of the formation (replacing gas) and
therefore the measured porosity values. The LWD average and maximum density
porosity values were increased mainly due to the increased fluid density. The wireline
acoustic porosity values were a poor match with core porosity values through the
Rotliegend gas interval possibly due to the high volume of gas (high porosity) affecting
the porosity values. Alternatively preferential acoustic energy propagation along low
porosity beds due to the dipping bedding encountered from the alternating dune slip
face/lee sands reduced the travel times. Note: Most open wells (uncased) become more
rugose and/or enlarged as time increases.

4-33
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for W ireline Standard model

100
— WAP Standard Frequency

WDP Standard Frequency

75
— WNP Standard Frequency

— CP Frequency (Adjusted)
g
| 50
fa

25

0
0 10 Porosity (%) 20

a)

Porosity histogram for LW D Standard model

100
— LMD P Standard Frequency

LAD P Standard Frequency

— LNP Standard Frequency

— CP Frequency (Adjusted)

o 10 Porosity (%) 20

b)

Figure 4.12 Vertical well model histograms with porosity ranging from 0 to 25% along
the x-axis and frequency 0 to 100 counts along the y-axis. Plot a) is the wireline derived
porosity standard distribution and plot b) is the LWD derived porosity standard
distribution. The black curve on all the plots represents the core porosity distribution.
The blue curve represents the wireline acoustic porosity distribution for plot a), but
represents the LWD average density porosity distribution for plot b). The red curve
represents the wireline acoustic porosity distribution for plots a), but represents the
LWD average density porosity distribution for plot b).

4-34
 Chapter 4: Analysis o f the Data Sets

Fluid P aram eters S tandard

Fluid Density (g/cm3) used for LW D Average Density log 1.000
Fluid Density (g/cm3) used for LW D M axim um Density log 1.000
Excavation correction (percentage porosity) used for LWD N eutron Porosity 0.000
Fluid Density (g/cm3) used for W ireline Density log 1.000
Fluid Slowness (ps/ft) used for W ireline Acoustic Porosity 189.0
Excavation correction (percentage porosity) used for W ireline N eutron Porosity 0.000

Table 4.17 Parameters used for porosity calculations from data set 1 vertical well logs
example.

Statistical tests at the LWD LWD

LW D W ireline W ireline W ireline
95% confidence level Average M axim um
N eutron Acoustic Density N eutron
(2 tailed) against Core Density Density
Porosity Porosity Porosity Porosity
Porosity Porosity Porosity

W hole X X ✓ ✓ ✓ X
Weissliegend X X X X ✓ X
Rotliegend Gas X X X ✓ ✓ X
Rotliegend W ater X ✓ ✓ ✓ ✓ ✓

Table 4.19 Vertical well statistical test results against core porosity values using the
models detailed above considered over intervals; Whole x008fit-x236ft, Weissliegend
x008ft-x 153ft, Rotliegend gas xl53ft-xl91ft and Rotliegend water xl91ft-x236ft.
X =reject the hypothesis and ^ a c c e p t the hypothesis at the 95% confidence level
using Chi2 goodness of fit test for grey cells and Student’s t-test with the hypothesised
mean difference=0 for clear cells.

C orrelation LWD LWD

LW D W ireline W ireline W ireline
coefficient against Average M axim um
N eutron Acoustic Density N eutron
C ore Porosity Density Density
S tandard 0.828 0.795 0.837 0.831 0.907 0.870

Table 4.20 Vertical well correlation coefficients against core porosity values over the
whole interval x008ft-x236ft.

4-35
 Chapter 4: Analysis o f the Data Sets

Data set 2 example

Well 1 wireline calculated and core porosity values were statistically compared
and the results of the linear shifted and Oil Company analyses are presented below
[Figure 4.17, Table 4.21 and Table 4.20]. The porosity histograms (0-40pu) are plotted
for the whole reservoir with well 1 core porosity values [Figure 4.17]. The standard
histogram includes the core, density, neutron, acoustic and the Oil Company density
and acoustic porosity values. The acronyms used were CPOR=well 1 core porosity,
DenPor=shale corrected density porosity, Nphi_S=sandstone neutron porosity,
AcouPor=shale corrected acoustic porosity, OC_DenPor=Oil Company shale corrected
density porosity and OC_AcouPor=Oil Company shale corrected acoustic porosity.
The first table detailing the Student’s t-tests for the clear cells and Chi2 tests for
grey cells, a tick/cross (V IX ) represents acceptance or rejection of the hypothesis at the
95% confidence level that there is zero difference between that porosity equation and
well 1 core porosity values over the fluid interval stated [Table 4.21]. The second table
gives details of the fluid parameters used in the equations for each fluid zone [Table
4.20]. In the fluid parameter table, OC refers to using the relevant Oil Company’s fluid
parmeter in that particular porosity equation. The oil zone is defined as oil at irreducible
water saturation. The transition zone is defined as oil at greater than irreducible water
saturation.
The statistical tests demonstrate that density and acoustic porosity values both
matched the core porosity values over each fluid zone, although the match was not good
when the comparison was made over the whole reservoir [Table 4.21]. The sandstone
neutron porosity values only match core porosity values in the oil zone. The Oil
Company’s (OC) density and acoustic porosity values only match core over the water
zone. The standard histogram shows that the OC’s fluid parameters [Table 4.20] led to
overestimation of the core porosity [Figure 4.17] whilst the density, neutron and
acoustic porosity equations match the core porosity values above 19% porosity. The
core porosity values around 10% porosity are due to high shale content in the core
samples in certain intervals. However, the removal of these values does not alter the
results presented. The acoustic and density values used to match with the core porosity
values and the OC’s values were quite different. The OC’s density value was derived
from the water density at hydrostatic pressure and the OC’s fluid travel time values was
that of water. However, the shallow investigation depth (~2inches) and the use of OBM

4-36
 Chapter 4: Analysis o f the Data Sets

to drill the well explains the use of 210psec/ft, the approximate fluid travel time for oil.
The fluid densities used reflect the change in fluid density within the formation. These
results show that selection of fluid parameters is important if porosity is to be estimated
with any accuracy. Interestingly, the OC’s density and acoustic porosity curves have
double peaks indicating the presence of the different zones [Figure 4.13].

Well 1: Standard Porosity Histogram

500
— C P O R (% )
— D enPor
— N p hi_S
400 — AcouPor
— O C_DenPor
O C AcouPor

S' 300
c
o
3
O’
2>
LL 200

100

0 10 20 30 40
Porosity (%)

F ig u re 4.13 Well 1 standard porosity histogram.

0.55 210 1.01 189

0.85 205 1.01 189
0.85 210 1.01 189
OC OC 1.01 189

T ab le 4 .2 0 Well 1 fluid parameters.

4-37
 Chapter 4: Analysis o f the Data Sets

✓ X ✓ X X
✓ ✓ ✓ ✓ X
✓ ✓ ✓ ✓ X
✓ ✓ ✓ ✓ ✓

X X X X X

Table 4.21 Well 1 Student’s t-tests results against well 1 core porosity values, fluid
parameters in Table 4.35. X =reject the hypothesis and /= accept the hypothesis at the
95% confidence level using Chi2 goodness of fit test for grey cells and Student’s t-test
with the hypothesised mean difference=0 for clear cells.

4-38
 Chapter 4: Analysis

4.3.3 Spectral analysis

Spectral «• “ * 1 ^ 4 31 J d D M M i « *

3 [Figure 4.6] and 6 [Figure . f series and provides the statistical

ttet . " ^ ^ r r u r i e r — of the autocorrelation frmction is the

significant frequencies in the ong

:" rrrrri
g transform of the data,
calculated by the inverse Fourier transform of the squmed ^ ^ ^
Autocorrelation is the process of summing a the
at increasing amounts of crossover. High auto—
significance level) occur when the resu t is * _ or
values may be a result of geological cyclicity ^ ^ ^ ^ ^ g y e d by comparison with
condition. Identification of the cause o any^pe ^ ^ ^ ^ the
autocorrelation function for other measurem ^cause they relate
autocorrelation functions of the rate of penetration „
t0 drilling effects and borehole conditions respectively. Other sign

assumed to result from geological features.

Figure 4.14 Schematic demonstrating a ‘horizontal’ on the

bed boundaries (black) several times, imparting a eye

logs.

4-39
 Chapter 4: Analysis o f the Data Sets

The log data were filtered to reduce the effect on the porosity values of non-
geological perturbations identified from the spectral analysis. The porosity analysis was
then repeated to test for improvement of the calculated porosity values. Filtering of the
log data was performed using a program that calculates the Fourier series of the input
log data to the number of required harmonics. The number of required harmonics
having been identified from the autocorrelation functions. The greater the number of
harmonics selected the greater the degree of accuracy of the match between the original
unfiltered log. The required harmonics were selected to filter the data for data set 1
(§4.4.1). Moving average filters were preferred for the data set 2 as this was less
damaging to geological features observed in the data (§4.4.2).

Example o f spectral analysis

An example of autocorrelation analyses of the wireline and LWD logs of the

data set 1 vertical well (x008-236ft) are presented in Figure 4.15. All the
autocorrelation plots in this chapter show the autocorrelation function for the
appropriate log (labelled at the head of each plot) and two coloured bands representing
the significance level of the autocorrelation function. If the autocorrelation function is
outside the yellow band, then the autocorrelation function is significant at that
wavelength at the 95% confidence level. If the autocorrelation function is outside the
blue and yellow bands, then the autocorrelation function is significant at that
wavelength at the 99% confidence level.

Autocorrelation for Vertical well Wireline Density Log

rO.01 (-) ■ r0.01 (+) r0.05 (+> r0.05 (-) WRHOB

^ 0.75 -
"ST

< 0.25 -

o 25 Distance (ft) 50

a)

4-40
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Vertical well Wireline Caliper Log

1 -

■ rO.Oi (-) ■ rO.01 (+) r0.05 (+) rO.05 <-) — WCALI

0.75 -
r-

0
I■5 0.5
0

1
< 0.25 -

0 -
o 25 Distance ( f t) 50

b)
Autocorrelation for Vertical well LWD Rate of Penetration Log
rO.01 (-) rO 01 (+) r0 .0 5 (+) r0.05 (-) LROP r(L)

0.75 -

0.25 ■

-0.25 -

-0.5
Distance (ft)

c)

Figure 4.15 Vertical well spectral analysis o f the wireline and LWD logs. Note: x-axis
is scaled from Distance 0 to 55ft, y-axis is scaled from Autocorrelation r(L) 0 to 1
above the x-axis. Values outside the yellow and blue bands are significant at the 0.05
and 0.01 confidence levels respectively.

The autocorrelation analyses o f the data set 1 vertical well wireline density (a),
caliper log (b) and LWD rate o f penetration log (c) indicate that there was little or no
effect on the density log by the borehole condition or drilling process [Figure 4.15].
This conclusion is based on the lack o f correlation between the autocorrelation
functions o f the wireline caliper (flat and characterless) and LWD rate o f penetration
(no features match with the density) and wireline density log. This approach is repeated
for all the logs for both data sets in section 4.4 below to identify if filtering is required.

4-41
 Chapter 4: Analysis o f the Data Sets

4.4 Analysis of the data sets

In the following section o f this chapter, the analyses o f both data sets are
presented. The porosity analyses o f the data sets included the use o f several different
porosity algorithms namely; standard formulae, linear shifted, square root, Wiley &
Patchett and hydrocarbon corrected using iterative calculations (§2.4.3 and §4.3). The
reservoirs were split into a number o f generic units based on the facies and fluid
changes from the vertical wells in each data set. Each porosity algorithm was applied to
each unit for both vertical and horizontal wells to assess the applicability o f the
algorithms with respect to the core porosity values from the vertical wells. The spectral
analyses o f the data sets enabled the evaluation o f contributing factors and the removal
o f some o f these unwanted influences (§4.3).

4.4.1 Data set 1

The reservoir was split into three generic units; Weissliegend x008- 153ft,
Rotliegend gas xl53-191ft and Rotliegend water x l 91-236ft for the vertical well
[Figure 4.1, layers marked] and Weissliegend xl34-906, xl655-1951ft and Rotliegend
gas x906-1655ft for the horizontal well [Figure 4.2, layers marked]. The idea was to
determine which method/s were sufficient to calculate accurate porosity values, since
there is no point in practice opting for a complex method over a simple method without
any notable improvement in accuracy and precision. It was also important to find
satisfactory LWD derived porosity to core porosity agreement because no wireline logs
for the horizontal well exist. Note that the vertical well was logged with LWD (LAD) 2
days AFTER the wireline tools.

Vertical well
Porosity analysis o f data set 1: vertical well

In the vertical well, it was possible to calculate porosity values from the density,
neutron and acoustic logs from both the wireline and LWD (no acoustic) tool
measurements. Figure 4.16, Table 4.23 and Table 4.24 contain the results o f the vertical
well standard porosity analyses. Table 4.22 lists the fluid parameters used.

4-42
 Chapter 4: Analysis o f the Data Sets

The porosity calculations clearly showed that wireline density porosity values
were most reliable. The wireline density porosity values were in good agreement with
core porosity values over all depth intervals using standard, single layer, three layer,
square root single layer, Wiley & Patchett standard, hydrocarbon corrected square root
and Wiley & Patchett models. The wireline density porosity values also had the highest
correlation with the core porosity values using the standard model.
The wireline neutron porosity values were in good agreement with core over all
depth intervals using single layer, square root single layer, Wiley & Patchett standard,
hydrocarbon corrected square root and Wiley & Patchett models. The wireline neutron
porosity values were not as accurate as the wireline density porosity values, but in
combination, their calculated porosity values were in good agreement with core.
The wireline acoustic porosity values were in poor agreement with core porosity
values only over the standard model Weissliegend, single layer Weissliegend and
Rotliegend gas and the three layer Rotliegend gas. The wireline acoustic porosity values
were a poor match with core porosity values through the Rotliegend gas interval
possibly due to the high volume o f gas affecting the porosity values. Alternatively
preferential acoustic energy propagation along low porosity beds due to the dipping
beds encountered from the alternating dune slip face/lee sands reduced the travel times.
3 Layer
Weiss­ Rotliegend Rotliegend
Standard Single Layer
liegend (Gas) (Water)
Fluid Density (g/cm3) used for
1.000 0.535 0.445 0.642 0.652
LWD Average Density log
Fluid Density (g/cm3) used for
1.000 0.649 0.543 0.781 0.796
LWD Maximum Density log
Excavation correction
(percentage porosity) used for 0.000 0.014 0.0188 0.004 0.005
LWD Neutron Porosity
Fluid Density (g/cm3) used for
1.000 0.947 0.927 0.972 0.977
Wireline Density log
Fluid Slowness (ps/ft) used for
189.0 184.0 182.3 185.7 186.3
Wireline Acoustic Porosity
Excavation correction
(percentage porosity) used for 0.000 0.029 0.038 0.009 0.011
Wireline Neutron Porosity

Table 4.22 Parameters used for porosity calculation from the logs.

4-43
 Chapter 4: Analysis o f the Data Sets

All the LWD porosity values were greatly improved by changes in fluid
parameters to be in good agreement with core porosity values over all the intervals
except for the LWD neutron porosity in the Rotliegend gas interval. The square root
method improved the LWD porosity values over the Wiley & Patchett method
especially with the average density values. Figure 4.16 showed that LWD porosity
generally over estimates core expect for the square root single and three layer models.

• The square root single model using the LWD average density and neutron porosity
values provides the best LWD derived porosity values in the vertical well.

Although accepted practice is that the LWD maximum density is preferred since
its density measurement is more valid and reliable with the LWD tools used (Rider
1996).
LWD neutron, average and maximum density porosity values were greater than
the wireline porosity values. The LWD neutron porosity values were increased with
respect to the wireline values due to the increased depth o f mud penetration increasing
the hydrogen index o f the formation (replacing gas). The LWD average and maximum
porosity values were increased mainly due to increase in formation fluid density
resulting from mud invasion. The fluid parameters used indicate the effect o f gas.
There was increased borehole rugosity/enlargement and thus worse formation to
detector contact than the wireline run indicated by the LWD caliper values. Note: Most
open wells (uncased) become more rugose and/or enlarged as time increases. Additional
evidence, in that the LWD density correction was negative whilst the wireline density
correction was positive indicated that the LWD density tool was correcting for the
heavy mud (barite). LWD PEF values were also lower than the wireline PEF values,
although if the barite were more significant at the time o f LWD logging, the expected
LWD PEF values would be greater than the wireline values. Note: barite collects at the
bottom o f the hole once mud circulation stops due to its high density.

4-44
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for W ireline Standard model

100
— WAP Standard Frequency

WDP Standard Frequency

75
— WNP Standard Frequency

— CP Frequency (Adjusted)
3 50

25

0
o 10 P orosity (% ) 20

a)

Porosity histogram for LWD Standard model

100
— LMD P Standard Frequency

LAD_P Standard Frequency

75
— LNP Standard Frequency

t? — CP Frequency (Adjusted)
g

25

0
o 10 Porosity (% ) 20

b)
Porosity histogram for Wireline 1 layer fluid model
100

— WAP 1 Layer Frequency

WDP 1 Layer Frequency -

— WNP 1 Layer Frequency

B 50
— CP Frequency (Adjusted)

0 10 P o ro sity (% ) 20

4-45
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for LWD 1 layer fluid model

100
— LADP 1 Layer Frequency

LMDP 1 Layer Frequency

— LNP 1 Layer Frequency

— CP Frequency (Adjusted)

0 10 Porosity (% ) 20

d)

Porosity histogram for Wireline 3 layer fluid model

100
— WAP_3_Layer Frequency

75 W D P 3 L a y e r Frequency -

— W N P3_Layer Frequency

2 .5 0
— CP Frequency (Adjusted)

25

0
0 10 P orosity (% ) 20

e)

Porosity histogram for LWD 3 layer fluid model

100
— LADP_3_Layer Frequency

LMDP_3_Layer Frequency

— LNP_3_Layer Frequency

— CP Frequency (Adjusted)
2 .50

o 10 P o ro s ity (•/•) 20

4-46
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for Square R oot Standard model

100
— LA D N SQ _P Standard Frequency

LMDN_SQ_P Standard Frequency

75
— WDN SQ P Standard Frequency

— CP Frequency (Adjusted)
5 .5 0

25

0
0 10 Porosity (•/•) 20

g)
P orosity histogram for W iiey & Patchett Standard model
100
— LADN_WP_P Standard Frequency

L M D N W P P Standa-d Frequency
75 -
— WDN WP P Standard Frequency

— CP Frequency (Adjusted)

50 -

25 -

o 10 Porosity (%) 20

h)

Porosity histogram for Square R oot 1 layer fluid model

100
— LADN_SQ_P 1 Layer Frequency

LMDN_SQ_P 1 Layer Frequency

— WDN_SQ_P 1 Layer Frequency

— CP Frequency (Adjusted)
5 .5 0 ■

25 -

o 10 P o ro s ity (% ) 20

i)

4-47
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for W iley & Patchett 1 layer fluid model

i
— LADN_WP_P 1 Layer Frequency

LMDN WP P 1 Layer Frequency

\f\\ — WDN_WP_P 1 Liiyer Frequency

f/\ \ \ — CP Frequency (Adtjusted)

/ / \ \ aA \ J N/n.

0 10 Porosity (% ) 20

j)
Porosity histogram for Square R oot 3 layer fluid model
100
— LADN_SQ_P_3_Layer Frequency

LMDN_SQ_P_3_Layer Frequency

— WDN_SQ_P_3_Layer Frequency

— CP Frequency (Adjusted)

3 50

0 10 Porosity (% ) 20

k)

Porosity histogram for W iley & Patchett 3 layer fluid model

100
— LADN_WP_P_3 Layer Frequency

LMDN_WP_P_3_Layer Frequency

— W D N W P P 3 _Layer Frequency

— CP Frequency (Adjusted)

2 .5 0

0 10 P o ro sity (% ) 20

l)

4-48
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for Hydrocarbon corrected Square Root fluid model

100
— L A D N S Q H C F req u en cy

LMDN_SQ_HC_Frequency

— WDN_SQ_HC_Frequency

CP Frequency (Adjusted)
■ 50-

0 10 Porosity (% ) 20

m)

Porosity histogram for Hydrocarbon corrected Wiley & Patchett fluid model
100
— L A D N W P H C F re q u ency

L M D N W P H C F re q u en c y
75
— WDN_WP_HC_Frequency

CP Frequency (Adjusted)

25

0
0 10 Porosity (%) 20

n)

Figure 4.16 Vertical well model histograms with porosity ranging from 0 to 25% along
the x-axis and frequency 0 to 100 counts along the y-axis. Plots a), c) and e) are
wireline derived porosity standard, single and three layer distributions respectively.
Plots b), d) and f) are LWD derived porosity standard, single and three layer
distributions respectively. The black curve on all the plots represents the core porosity
distribution. The blue curve represents the wireline acoustic porosity distribution for
plots a), c) and e); the LWD average density porosity distribution for plots b), d) and f);
and the resulting LWD average density and neutron porosity distribution for plots g) to
n). The red curve represents the wireline acoustic porosity distribution for plots a), c)
and e); the LWD average density porosity distribution for plots b), d) and f); and the
resulting LWD average density and neutron porosity distribution for plots g) to n).

4-49
 II
Chapter 4: Analysis o f the Data Sets

Statistical tests at the LWD LW D

LWD W ireline W ireline W ireline
95% confidence level Average M aximum
N eutron Acoustic Density Neutron
(2 tailed) against Core Density Density
Porosity Porosity Porosity Porosity
Porosity Porosity Porosity

W hole X X ✓ ✓ ✓ X
Weissliegend X X X X ✓ X
Rotliegend Gas X X X ✓ ✓ X
Rotliegend W ater X ✓ ✓ ✓ ✓ ✓

W hole ✓ ✓ ✓ ✓ ✓ ✓
Weissliegend ✓ ✓ ✓ X ✓ ✓
Rotliegend Gas ✓ X X X ✓ ✓
Rotliegend W ater ✓ ✓ ✓ ✓ ✓ ✓

W hole ✓ ✓ ✓ ✓ ✓ ✓
Weissliegend ✓ ✓ ✓ ✓ ✓ ✓
Rotliegend Gas ✓ ✓ X X ✓ X
Rotliegend W ater ✓ ✓ ✓ ✓ ✓ ✓
Using LW D Average Using LW D M aximum Using W ireline
Density and N eutron Density and Neutron Density and N eutron

Whole ✓ ✓ ✓
Weissliegend X ✓ X
Rotliegend Gas ✓ ✓ X
Rotliegend W ater ✓ ✓ ✓

W hole ✓ ✓ ✓
W eissliegend ✓ ✓ ✓
Rotliegend Gas ✓ X ✓
Rotliegend W ater ✓ ✓ ✓

Whole ✓ ✓ ✓
W eissliegend ✓ ✓ ✓
Rotliegend Gas ✓ X X
Rotliegend W ater ✓ ✓ ✓

4-50
 Chapter 4: Analysis o f the Data Sets

Whole X X ✓
Weissliegend X X ✓
Rotliegend Gas X X ✓
Rotliegend W ater X X ✓

Whole ✓ ✓ ✓
Weissliegend X X X
Rotliegend Gas ✓ ✓ X
Rotliegend W ater ✓ ✓ ✓

Whole ✓ ✓ ✓
W eissliegend X X X
Rotliegend Gas ✓ ✓ ✓
Rotliegend W ater ✓ ✓ ✓
Using LWD
Using LW D Average Using W ireline
M axim um Density
Density and N eutron Density and Neutron
and N eutron

W hole ✓ ✓ ✓
W eissliegend X X ✓
Rotliegend Gas ✓ ✓ ✓
Rotliegend W ater ✓ ✓ ✓

Whole ✓ ✓ ✓
W eissliegend X X ✓
Rotliegend Gas X ✓ ✓
Rotliegend W ater X ✓ ✓

Table 4.23 Vertical well statistical test results against core porosity values using the
models detailed above considered over intervals; Whole x008ft-x236ft, Weissliegend
x008ft-x 153ft, Rotliegend gas xl53 ft-x l9 1 ft and Rotliegend water xl91ft-x236ft.
X =reject the hypothesis and /= a c c e p t the hypothesis at the 95% confidence level

using Chi2 goodness o f fit test for the grey cells and Student’s t-test with the
hypothesised mean difference=0 for the clear cells.

4-51
 Chapter 4: Analysis o f the Data Sets

Correlation LWD LWD

LWD W ireline W ireline W ireline
coefficient against Average M aximum
Neutron Acoustic Density Neutron
Core Porosity Density Density
Standard 0.828 0.795 0.837 0.831 0.907 0.870
Single Layer 0.828 0.795 0.837 0.907 0.831 0.870
3 Layer 0.850 0.832 0.804 0.906 0.840 0.798
Using LW D Average Using LWD M aximum Using W ireline Density
Density and N eutron Density and Neutron and Neutron

Standard 0.859 0.844 0.876

Single Layer 0.863 0.852 0.881
3 Layer 0.863 0.853 0.851

Standard 0.862 0.846 0.876

Single Layer 0.864 0.850 0.877
3 Layer 0.865 0.853 0.850

Square Root 0.756 0.742 0.833

Wiley & Patchett 0.865 0.850 0.881

T able 4.24 Vertical well correlation coefficients against core porosity values over the
whole interval x008ft-x236ft.

4-52
 Chapter 4: Analysis o f the Data Sets

Spectral analysis of data set 1: vertical well

Autocorrelation analyses o f the wireline and LWD logs through the reservoir in
the vertical well (x008-236ft) are presented below [Figure 4.17]. The plots all show the
autocorrelation function for the appropriate log and the significance level o f the
autocorrelation function (§4.3.3). The autocorrelation analysis is summarised below in
Table 4.25 giving the significant wavelengths at the 99% confidence level.
The autocorrelation analyses o f the vertical well wireline logs indicated that the
logs appear to be largely unaffected by the borehole or drilling process i.e. there is a
lack o f any correlation between the caliper and LWD rate o f penetration autocorrelation
functions and that o f any o f the other logs [Table 4.25 and Figure 4.17]. The only
exception was the wireline PE F log with wavelengths at 30-32ft, 45-48ft and 50-52fl
which correlated with the LWD caliper log [Table 4.25 and Figure 4.17c and q]. The
similarities were unlikely to be coincidental and m ay have been related to drilling
induced spiralling o f the vertical well (see examples in §3.3.2). This was especially
evident in the LWD caliper which displayed resonance features (repeated significant
peaks at equal intervals). The fact that the LWD density correction log was also
affected explained why the LWD photoelectric factor and density logs were not affected
in the same manner. The LWD density correction had removed these features from the
logs and thus removed the effect o f the spirals that may have been re-exposed by the
drillstring during the LWD logging. However, the same wavelengths did not appear to
be affected for the wireline caliper log which was almost constant in value i.e. the
borehole was smooth [Figure 4.1 and Figure 4.17q]. Alternatively, mudcake may have
filled the spirals, or tool wobble, or preferential mudcake build up on the more porous
facies (dune slip faces), which was subsequently measured by the shallow reading PEF
log.

• The vertical well wireline and LWD data was not filtered because only marginal
improvement in the accuracy o f the porosity calculations would be gained.

4-53
 Chapter 4: Analysis o f the Data Sets

Below 7ft Below 5ft

Gamma ray a) and b)
16-20ft 8-14ft
Below 5 ft
23-27ft
Below 22ft
Photoelectric factor 30-32ft c) and d)
Above 36ft
45-48ft
50-52ft
Deep induction e)
All All
Amplitude resistivity f)
Medium induction g)
All All
Phase resistivity h)
Density 0
Maximum density Below 42ft Below 52ft j)
Average density k)
Acoustic travel time All 1)
Neutron porosity All All m) and n)
Below 5 ft
18-26ft
33-34ft
Density correction Below 2ft o) and p)
3 8-39ft
43-49ft
50-51ft
Below 2ft
6-8ft
12-16ft
21-23ft
Caliper All 26-31ft q) and r)
32-33ft
35-38ft
45-46ft
48-52ft
Below 24ft
33-38ft s)
Above 42ft

Table 4.25 Data set 1 vertical well autocorrelation analysis giving the significant
wavelengths at the 99% confidence level.

4-54
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Vertical well Wireline Gamma Ray Log

r 0 .0 1 (-) r0.01 (+) r0.05 (+) r0.05 (-) WGR

0.75

0.5

0.25

50

-0.25 Distance (ft)

a)
Autocorrelation for Vertical well LWD Gamma Ray Log
r0.01 (-) r0.01 (+) rO.05 (+) r0.05 (-) L G R r(L )

0.75 -
-I
"nr
c
0 0.5 -

1
fc
8 0- 2 5 -
1
<

-0.25
Distance (ft)

b)
Autocorrelation fo r Vertical w ell W ireline Photoelectric factor Log
r0.01 (-) r0.01 (+) rO.05 (+) r0.05 (-) W PEF

0.75

O 0.5

8 0.25

50

-0.25
Distance (ft)

4-55
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Vertical w ell LWD Photoelectric factor Log

r0.01 (-) r0.01 <+) r0.05 (+) r0.05 (-) LPEF r(L)

0.75

0.5

•S 0.25

-0.25

-0.5
Distance (ft)

d)

Autocorrelation for Vertical well Wireline Deep Induction Log

■ r0.01 (-) ■ r0.01 (+) r0.05 (+) r0.05 (-) WILD

■a 0.5 -

< 0 .2 5 -

o 25 Distance (ft) 50

e)
Autocorrelation for Vertical well LWD Am plitude Resistivity Log
1-0.01 (-) r0.01 (+) r0.05 (+) r-0.05 (-) LARes r(L)

< 0 .2 5 -

o 25 Distance (ft) 50

4-56
 Chapter 4: Analysis o f the Data Sets

Autocorrelation fo r Vertical well W ireline Medium Induction Log

r0.01 (-) r0.01 (+) r0.05 (+) r0.05 (-) W1LM

0.75 -

0.5 -

< 0.25 -

0 25 Distance (ft) 50

g)
Autocorrelation for Vertical well LWD Phase Resistivity Log
f0.01 (-) r0.01 <+) r0.05 (+) r0.05 (-) L P R es r(L)

'-J0 .7 5 -

“S 0.5 -

< 0 .2 5 -

o 25 Distance (ft) 50

h)

Autocorrelation for Vertical well Wireline Density Log

r0.01 (-) r0.01 (+) r0.05 (+) r0.05 (-) WRHOB

■0.75 -

0 .2 5 -

o 25 Distance (ft) 50

4-57
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Vertical well LWD Maximum Density Log

1
r0.01 (-) r0.01 (+) r0.05 (+) r0.05 (-) LMRhob r<L)

"35 0.5

< 0 .2 5

0
o 25 Distance (ft) 50

j)

1
Autocorrelation for Vertical well LWD Average Density Log
■ r0.01 (-) B r0.01 (+) r0.05 (+) r0.05 (-) LARhob r(L)

35 0.5

< 0 .2 5

0
o 25 Distance (ft) 50

k)

Autocorrelation for Vertical well W ireline Acoustic Travel Time Log

r0.01 (-) r0.01 (+) K3.05 (+) r0.05 (-) WDT

"35 0.5

< 0 .2 5

o 25 Distance (ft) 50

l)

4-58
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Vertical well W ireline Neutron Porosity Log

r0.01 r0.01 (+) r0.05 (+) r0.05 (-) WNPHI

■S 0.5

< 0 .2 5

o 25 Distance (ft) 50

m)

Autocorrelation for Vertical well LWD Neutron Porosity Log

r0.01 (-) r0.01 (+) r0.05 (+) r0.05 ( - ) -------LNPHI r<L)

0.75 -

0.5 -

O
8
< 0.25 -

o 25 Distance (ft) 50

n)

Autocorrelation for Vertical well W ireline Density Correction Log

r0.01 (-) r0.01 (+) rO 05 (+) r0.05 (-) WDRHO

0.75

0.2 5 -

50

-0.25
Distance (ft)

4-59
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Vertical well LWD Density Correction Log___

* r0.01 (-) r0.01 (+) r0.05 (+) r0.05 (-)— » LDRHO r(L)

0.75

0.5

75 0.25

-0.25

-0.5
Distance (ft)

1 -
Autocorrelation for Vertical well Wireline Caliper Log

m r0.01 (-) * r0.01 (+) r0.05 (+) r0.05 (-) ----- WCALI

0.75 -
r-
c
0
X3
■
g 0 .5 -

1
< 0.25 -

0 -I i I
o 25 Distance (ft) 50

Autocorrelation for Vertical well LWD Caliper Log

r0.01 (-) Sc r0.01 (+) rO.05 (+) LCALI r<L)

0.75

0.2 5 -

-0.25
Distance (ft)
 Chapter 4: Analysis o f the Data Sets

Autocorrelation fo r Vertical well LWD Rate of Penetration Log

r0.01 (-) r0.01 (+) r0.05 (+) r0.06 (-) LROP r<L)

0.75

V 0.5

■s 0 25

-0.25

-0.5
Distance (ft)

s)

Figure 4.17 Vertical well spectral analysis o f the wireline and LWD logs. Note: x-axis
is scaled from Distance 0 to 55ft, y-axis is scaled from Autocorrelation r(L) 0 to 1
above the x-axis. Values outside the yellow and blue bands are significant at the 0.05
and 0.01 confidence levels respectively.

4-61
 Chapter 4: Analysis o f the Data Sets

Horizontal well

In the horizontal well, porosity values were calculated from the LWD density
and neutron tool measurements. The same algorithms for calculating porosity were used
as for the vertical well, but without the benefit o f horizontal well core porosity data to
verify the LWD calculated porosity values. In addition, only two o f the three facies
intervals were observed in the horizontal well; Weissliegend xl34-906ft and xl655-
1951ft, and Rotliegend gas x906-1655ft [Figure 4.2, layers marked].
The horizontal well LWD neutron and maximum density porosity values were
lower than their respective measurements in the vertical well. The LWD neutron
porosity values were lower than the vertical well LWD neutron porosity values, but
comparable with the wireline neutron porosity values. The horizontal well LWD
measurements were recorded approximately 18 hours after penetration and the vertical
well wireline measurements 36 hours after penetration. The volume o f mud invasion
was similar in both cases when compared with the volume o f investigation o f the
porosity tools. The vertical well LWD neutron porosity measurements were recorded
approximately 84 hours after penetration and therefore mud invasion was more
extensive, but also allowed gas migration towards the borehole.
The horizontal well LWD average density values were very low and the LWD
caliper values were large. This suggested that the borehole was enlarged and rugose.
The magnitudes o f the LWD average density and caliper values were large enough to
mean that both measurements were unreliable for any quantitative use and cast doubts
on the reliability o f all the other horizontal well LWD measurements.

Porosity analysis o f data set 1: horizontal well

The lack o f core porosity data made it difficult to determine which method(s)
were appropriate for calculating accurate horizontal well porosity values. However,
comparison o f LWD log calculated porosity with the scaled vertical well core porosity
values over the relevant facies intervals allowed rejection o f methods which were
certainly inaccurate. Correlation was not possible between the vertical well core
porosity and horizontal well calculated porosity values. A list o f the fluid parameters for
all the methods used are given below [Table 4.26].

4-62
 Chapter 4: Analysis o f the Data Sets

2 Layer
Standard Single Layer Weissliegend Rotliegend (Gas)
Fluid Density (g/cm3) used for
1.000 0.535 0.445. 0.642
LWD Average Density log
Fluid Density (g/cm3) used for
1.000 0.649 0.543 0.781
LWD Maximum Density log
Excavation correction
(percentage porosity) used for 0.000 0.014 0.0188 0.004
LWD Neutron Porosity

Table 4.26 Parameters used for porosity calculation from the horizontal well logs.

Horizontal well porosity calculations were performed in the same manner as for
the vertical well above. The results o f the porosity analyses are presented below [Figure
4.18 and Table 4.27]. The horizontal well LWD maximum density and neutron porosity
values were in good agreement with core porosity values over the whole and
Weissliegend depth intervals using Wiley & Patchett standard, single layer and two
layer models. It was possible using the LWD average density porosity values to achieve
good agreement with core porosity values over some depth intervals using some of the
models. Although the corrections applied were large and unreliable due to the spread o f
the LWD average density porosity values.

• The standard Wiley & Patchett LWD maximum density porosity values are most
likely to provide the best LWD derived porosity values in the horizontal well.

This was not the same model as the vertical well (standard square root model)
suggesting that reliable and consistent porosity calculation from these LWD porosity
tools was not viable in this well or possibly the whole field.
The horizontal well LWD maximum density porosity values were significantly
lower than both the vertical wireline density and LWD maximum density porosity
values. This was possibly due to barite mud invasion and mudcake increasing
concentration o f barite close to the borehole walls in the formation and mudcake at the
time o f logging compared with the vertical well. This would effectively have increased
the formation fluid density to greater than lg/cm3 reducing the calculated porosity
values. The increased effect o f the barite was indicated by the very large LWD PEF

4-63
 Chapter 4: Analysis o f the Data Sets

values in the horizontal well. The PEF measurement is only sensitive to the first V&inch
in front o f the short spaced detector indicating that the barite concentration was much
greater than in vertical well measurements and presumably lying on the borehole floor.
An alternative scenario was the incorrect LWD tool position within the borehole
whilst logging. A known phenomenon is the LWD tool “riding the borehole” in which
the tool climbs the borehole wall in the direction o f drilling (clockwise) which leads to
the possibility o f non-parallel standoff (§3.3.1). This non-parallel standoff can lead to
either light or heavy bulk density values with no indication from the density correction
curve. This situation can be understood if one considers calculating the density from
each detector separately. The bulk density is a combination o f these two densities,
which can be approximated by:

4 1
Ph 3 Pi 3 Ps

where pb = bulk density, pi = long spaced density and p s = short spaced density
(Sherman and Locke 1975). Thus, in a situation where the short spaced detector reads
light (poor contact) and the long spaced detector reads correctly (good contact) then the
bulk density would read too heavy. This situation is possible since the long spaced
detector leads the short spaced detector and the source down the borehole, the reverse
o f a wireline tool.

4-64
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for LW D Standard model

1000
— LAD_P Standard Frequency

— LMD_P Standard Frequency

750 — LNP Standard Frequency

— CP Frequency (Adjusted)

500

250

0 10 Porosity (%) 20 30

a)
Porosity histogram for LW D 1 layer fluid model
1000
— LADP 1 Layer Frequency

LMDP 1 Layer Frequency

750
— LNP 1 Layer Frequency

■CP Frequency (Adjusted)

3 500

250

0 10 Porosity (%) 20 30

b)

Porosity histogram for LW D 2 layer fluid model

1000
— LADP_2_Layer Frequency

LMDP_2_Layer Frequency
750
— LNP_2_Layer Frequency

CP Frequency (Adjusted)
2 .5 0 0

250

0 10 P oro sity (% ) 20 30

c)

4-65
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for Square Root Standard model

1000 r
— LADN_SQ_P Standarc [Frequency
LMDN_SQ_P Standar d Frequency
— CP Frequency (Adjust)

fc*
S
3 500 N I/
/
t
fi

250
f \ / ~ \ JV V.
/ ^ \\ /
/ ' —v
\
J y
0 10 Porosity (%) 2.0 0

d)

Porosity histogi•am for W iley & Patelliett Standard model

1000
— LADN_WP_P Standard Frequency
LMDN_WP_P Standard Frequency
— CP Frequency (Adjuste< !)
750

e>
3 500
I I
C*< / 1 / \

250

10 Porosity (% ) 20 30

e)

Porosity histogram for Square R oot 1 layer fluid model

1000
— LADN SQ P 1 Layer Frequency
LMDN_SQ_P 1 Layer Frequency
— CP Frequency (Adjusted)
750

g.500

250

0 10 P o ro s ity (% ) 20 30

f)

4-66
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for W iley & Patchett 1 layer fluid model

1000
— L A D N W P P 1 Layer Frequency
LMDN_WP_P 1 Layer Frequency
CP Frequency (Adjusted)

3 500

10 Porosity (%) 20

g)
Porosity histogram for Square R oot 2 layer fluid model
1000
— LADN_SQ_P_2_Layer Frequency
LMDN_SQ_P_2_Layer Frequency
— CP Frequency (Acfousted)
750

3 500

250

0 10 Porosity (% ) 20 30

h)

Porosity histogram for W iley & Patchett 2 layer fluid model

1000
— LADN WP P 2_Layer Frequency
LMDN_WP_P_2_Layer Frequency
— CP Frequency (Adjusted)
750

500

250

0 10 P oro sity (% ) 20 30

i)

4-67
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for Hydrocarbon corrected Square Root fluid model

1000
— LADN_SQ_HC_Frequency
LMDN_SQ_HC_Frequency
— CP Frequency (Adjusted)
750

2 .5 0 0

250

0 10 Porosity (% ) 20 30

j)
Porosity histogram for Hydrocarbon corrected Wiley & Patchett fluid model
1000
— LADN_WP_HC_Frequency
L M D N W P H C F requency
— CP Frequency (Adjusted)
750

250

0
o 10 Porosity (%) 20 30

k)

Figure 4.18 Horizontal well model histograms with porosity ranging from 0 to 35%
along the x-axis and frequency 0 to 1000 counts along the y-axis. The black curve on all
the plots represents the core porosity distribution. The blue curve represents the LWD
average density porosity distribution for plots a) to c), and the resulting LWD average
density and neutron porosity distribution for plots d) to k). The red curve represents the
LWD maximum density porosity distribution for plots a) to c). The green curve
represents the LWD neutron porosity for plots a) to c), and the resulting LWD average
density and neutron porosity distribution for plots d) to k).

4-68
 Chapter 4: Analysis o f the Data Sets

Statistical tests at the 95% confidence LWD Average LW D M aximum LW D Neutron

level (2 tailed) against C ore Porosity Density Porosity Density Porosity Porosity

Whole X ✓ ✓
Weissliegend X X X
Rotliegend Gas X X X

Whole X X ✓
Weissliegend X X ✓
Rotliegend Gas ✓ X X

Whole X X ✓
Weissliegend ✓ X ✓
Rotliegend Gas X X X
Student’s t-test at the 95% confidence Using LWD Average Using LW D M aximum
level (2 tailed) against C ore Porosity Density and N eutron Density and N eutron

Whole X ✓
Weissliegend X X
Rotliegend Gas ✓ X

W hole / ✓
Weissliegend X ✓
Rotliegend Gas X X

Whole ✓ ✓
Weissliegend X ✓
Rotliegend Gas ✓ X

Whole X ✓
Weissliegend X ✓
Rotliegend Gas X X

Whole
Weissliegend
Rotliegend Gas

4-69
 Chapter 4: Analysis o f the Data Sets

Whole ✓ ✓
Weissliegend X ✓
Rotliegend Gas ✓ X

Whole X ✓
Weissliegend X X
Rotliegend Gas ✓ X

Whole X ✓
Weissliegend X X
Rotliegend Gas X X

Table 4.27 Horizontal well statistical test results against core porosity considered over
intervals; Whole xl34ft-xl951ft, Weissliegend xl34ft-x906, xl655-1951ft and
Rotliegend gas x906ft-xl951ft. X =reject the hypothesis and y = a c c e p t the hypothesis
at the 95% confidence level using Chi2 goodness o f fit test for the grey cells and
Student’s t-test with the hypothesised mean difference=0 for the clear cells.

4-70
 Chapter 4: Analysis o f the Data Sets

Spectral analysis o f data set 1: horizontal well

The autocorrelation analysis o f the horizontal well ( x l34-195lft) LWD logs

suggested that most o f the logs were affected by drilling or the borehole due to the
correlation o f the caliper with the other logs [Figure 4.19 and Figure 4.2]. Only the
LWD resistivity logs were unaffected. The autocorrelation analyses are summarised
below in Table 4.28 giving the significant wavelengths at the 99% confidence level.
The LWD density correction algorithm appeared to be unable to correct
sufficiently for the standoff encountered and spiralling is evident at a number o f
wavelengths in LWD caliper log. O f particular note is the LWD caliper autocorrelation
functions with significant wavelength features <180ft, 210-225ft, 270-345ft and 370-
385ft [Figure 4.19h]. The most damaging wavelengths appear to be the short
wavelengths between 8-10ft. Consequently, the logs were filtered using the first 180
harmonics, which removed all wavelengths below 10ft inclusively. The porosity
calculations were then repeated on the filtered horizontal well LWD data.

Below 180ft
Gamma ray a)
Above 260ft
Below 280ft
Photoelectric factor 310-340ft, 345-370 b)
390-395,445-450
Amplitude resistivity
All c)
Phase resistivity
Maximum density Maximum Below 395ft d)
Average density (Average all) e)
Neutron porosity All f)
Below 210ft
Density correction g)
Above270ft
Below 180ft, 210-225ft
Caliper h)
270-345ft, 370-385ft
Below 340ft, 350-365ft
Rate of penetration i)
Above 375ft

Table 4.28 Data set 1 horizontal well autocorrelation analysis giving the significant
wavelengths at the 99% confidence level.

4-71
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Horizontal well LWD Gamma Ray Log

r0.01 (-) r0.01 r0.05 rO.Q5 (-) LGR r(L)

0.75

0.5

0.25

100 200

-0.25
Distance (ft)

a)
Autocorrelation fo r Horizontal w ell LWD Photoelectric factor Log
r0.01 (-) r0.01 r0.05 (-) LPEF r(L)

0.75

0.5

0.25

<
100 200 300 400

-0.25
Distance (ft)

b)
Autocorrelation for Horizontal well LWD Resistivity Log
1
r0.01 (-) r0.05 r0.05 (-) L R es r(L)

_ |0 .7 5

0) 0.5

< 0 .2 5

0
0 100 200 300 400
Distance (ft)

c)

4-72
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Horizontal well LWD Maximum Density Log

r € .0 1 (-) r0.01 r0.05 r 0 .0 5 (-) LMRhob r<L)

0.75 -

S 0-25 -

100 200 300 400

-0.25
Distance (ft)

d)
Autocorrelation fo r Horizontal well LWD Average Density Log
1
r0.01 (-) r0.01 r0.05 rO.OS (-) LARhob r(L)

- J 0 .7 5

ai 0.5

0 .25

0
0 100 200 300 400
Distance (ft)

e)
Autocorrelation for Horizontal well LWD Neutron Porosity Log

r0.01 (-) r 0 .0 1 r0.05 r0.05 (-) LNPHI r(L)

- J 0 .7 5
•cr

0 .25 -

0 100 200 300 400

Distance (ft)

f)

4-73
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Horizontal well LWD Density Correction Log

r0.01 (-) r0.01 r0 .0 5 rO.Q5 (-) LDRHO r(L)

0 .7 5 -

8 025

100 200 400

-0.25
Distance (ft)

g)
Autocorrelation for Horizontal well LWD Caliper Log
r0.01 (-) r0.01 r0.05 10.05 (-) LCALI r(L)

0 .7 5 -

0 .2 5 -

100 200

-0 .2 5
Distance (ft)

h)
A u to c o rre la tio n fo r H orizontal w ell l WD R ate o f P e n e tra tio n Log
r0.01(-) r0.01 r0.05 r0 .0 5 (-) LROP r<L)

« 0 .5 -

< 0 .2 5 -

0 100 200 300 400

Distance (ft)

i)

Figure 4.19 Horizontal well spectral analysis o f the LWD logs. Note: x-axis is scaled
from Distance 0 to 450ft, y-axis is scaled from Autocorrelation r(L) 0 to 1 above the x-
axis. Values outside the yellow band are significant at the 0.05 confidence level and at
the 0.01 confidence level outside the blue band.

4-74
 Chapter 4: Analysis o f the Data Sets

Porosity analysis o f data set 1: horizontal well filtered data

The filtered horizontal well LWD data were used to calculate porosity values.
The same approaches for calculating porosity were used as for the unfiltered data. The
additional knowledge o f the mud invasion was taken into account and fluid densities
greater than lg/cm3 were used. Comparisons between the filtered LWD log calculated
porosity with the scaled vertical well core porosity values allowed rejection o f methods,
which were inaccurate i.e. poor match with the vertical well core porosity values.
Initially, the standard model was used to estimate formation porosity in the horizontal
well. A list o f the fluid parameters for all the methods used are given [Table 4.29].
The horizontal well filtered porosity analyses results are given below [Figure
4.20 and Table 4.30]. The horizontal well filtered porosity values were poor using the
standard model and were significantly improved by alteration o f the fluid parameters.
LWD maximum density values were in good agreement with core porosity values over
all depth intervals using single layer and single layer Wiley & Patchett standard models.
The horizontal well filtered LWD average density and neutron porosity values were
much improved with respect to core porosity values over the whole and Weissliegend
intervals using the single and two layer models. The horizontal well filtered data LWD
porosity values used in were much improved with respect with core porosity values
over most intervals [Figure 4.20 and Table 4.30].

• The best porosity values in the horizontal well were provided by the filtered LWD
maximum density and neutron porosity single layer Wiley & Patchett model.

2 Layer
Standard Single Layer Weissliegend Rotliegend (Gas)
Fluid Density (g/cm3) used for
1.000 -0.531 -0.630 -0.419
LWD Average Density log
Fluid Density (g/cm3) used for
1.000 1.078 0.810 1.236
LWD Maximum Density log
Excavation correction
(percentage porosity) used for 0.000 0.020 0.026 0.010
LWD Neutron Porosity

Table 4.29 Parameters used for porosity calculation from the horizontal well logs.

4-75
 Chapter 4: Analysis o f the Data Sets

Porosity- h is to g ra m f o r L W D S ta n d a r d m odel
1000
— LARHOB Standard
— LMRHOB Standard
LNPHI Standard
— CP (Adjusted)
750

3.500

250

0 10 Porosity (%) 20 30

a)

Porosity h istogram for LW D 1 1:aver fluid model

— LARHOB Single Layer
LMRHOB Single Layer
— LNPHI Single Layer
— CP (Adjusted)

k / \
4 V \
i K \
\ \
i/ » \ \

rf

0 10 Porosity (%) 20 30

b)

1000
Porosity histogram for LW D 2 layer fluid model
— LARHOB Two Layer
LMRHOB Two Layer
— LNPHI Two Layer
— CP (Adjusted!)
750

2.500

250

0 10 P o ro s ity (% ) 20 30

4-76
 Chapter 4: Analysis o f the Data Sets

1000
— LARHOB+NPHI Square Itoot Standard
LMRHOB+NPHI Square Root Standard
— CP (Adjusted)
750

fe*
s3 500
t
m
\ r--- \
250

.. A l . j
10 Porosity (% ) 20 30

d)
Porosity histogram for W iley & Patchett Standard model
1000
— LARHOB+NPHI Wiley & Patchett Standard
LMRHOB+NPHI Wiley & Patchett Standard
— CP (Adjusted)
750

3 500

250

0 10 Porosity (%) 20 30

e)
Porosity histogram for Square R oot 1 layer fluid model
1000
— LARHOB+NPHI Square Root Single Layer
LMRHOB+NPHI Square Root Single Layer
— CP (Adjusted)
750

3 500

250

0 10 P o ro sity (% ) 20 30

4-77
 Chapter 4: Analysis o f the Data Sets

1000 Porosity histogram for W iley & Patchett 1 layer fluid model
— LARHOB+NPHI Wiley & Patchett Single Layer
LMRHOB+NPHI Wiley & Patchett Single Layer
^ C P (Adjusted)

3 500

10 Porosity (%) 20

g)

1000
Porosity histogram for Square R oot 2 layer fluid model
— LARHOB+NPHI Square Root Two Layer
LMRHOB+NPHI Square Root Two Layer
— CP (Adjusted)

10 Porosity (%) 20

h)

1000 for W iley & Patchett 2 layer fluid model

— LARHOB+NPHI Wiley & Patchett Two Layer
LMRHOB+NPHI Wiley & Patchett Two Layer
— CP (Adjusted)
750

9 500

250

0 10 P o ro sity (%) 20 30

i)

4-78
 Chapter 4: Analysis o f the Data Sets

Porosity histogram for Hydrocarbon corrected Square Root fluid model

1000
— LARHOB+NPHI HC SqPor
LMRHOB+NPHI HC SqPor
— CP (Adjusted)
750

2 .5 0 0

250

0 10 Porosity (% ) 20 30

j)
Porosity histogram for Hydrocarbon corrected Wiley & Patchett fluid model
1000 ■................................. !i 1 ............. n
— LARHOB+>IPHI HC WPPor
LMRHOB+1■JPHI HC WPPor
— CP (Adjusts fl
750 I

|
3 ,5 0 0

£t

250 i
\
>

10 Porosity (%) 20 30

k)

Figure 4.20 Horizontal well model histograms using filtered data with porosity ranging
from 0 to 35% along the x-axis and frequency 0 to 1000 counts along the y-axis. The
black curve on all the plots represents the core porosity distribution. The blue curve
represents the LWD average density porosity distribution for plots a) to c), and the
resulting LWD average density and neutron porosity distribution for plots d) to k). The
red curve represents the LWD maximum density porosity distribution for plots a) to c).
The green curve represents the LWD neutron porosity for plots a) to c), and the
resulting LWD average density and neutron porosity distribution for plots d) to k).
Note: the filtering was performed to removing the unwanted harmonics.

4-79
 Chapter 4: Analysis o f the Data Sets

S tudent’s t-test at the 95% confidence L W D A v e rag e L W D M a x im u m LW D Neutron

level (2 tailed) against Core Porosity D e n s ity P o r o s it y D e n s ity P o r o s it y Porosity

Whole
Weissliegend
Rotliegend Gas
Single Layer
Whole
Weissliegend
Rotliegend Gas

Whole
Weissliegend
Rotliegend Gas
Using LW D Average Density Using LW D M aximum
and N eutron Density and N eutron
S quare Root
S tandard
Whole
Weissliegend
Rotliegend Gas
Single Layer
Whole
Weissliegend
Rotliegend Gas

Whole
Weissliegend
Rotliegend Gas
Wiley & Patchett
S tandard
W hole
Weissliegend
Rotliegend Gas
Single Layer
Whole
W eissliegend
Rotliegend Gas

4-80
 Chapter 4: Analysis o f the Data Sets

Whole ✓ ✓
Weissliegend X X
Rotliegend Gas X ✓

W hole X ✓
Weissliegend X X
Rotliegend Gas X ✓

W hole X X
Weissliegend ✓ V
Rotliegend Gas X X

Table 4.30 Horizontal well filtered data statistical test results against core porosity
considered over intervals; Whole x l3 4 ft-xl951ft, W eissliegend xl34ft-x906, xl655-
1951ft and Rotliegend gas x906ft-xl951ft. X =reject the hypothesis and </ =accept the
hypothesis at the 95% confidence level using Chi2 goodness o f fit test for grey cells and
Student’s t-test with the hypothesised mean difference=0 for clear cells.

4-81
 Chapter 4: Analysis o f the Data Sets

4.4.2 Data Set 2

For data set 2 the reservoir was subdivided into four units based on the change
in formation fluid from gas to oil to water with a transition zone between oil and water
zones [Table 4.31]. The porosity analyses were performed on the whole reservoir and
the four units separately for each o f the porosity algorithms in turn. Calculating porosity
in the shaley mixed-fluid sandstone reservoir, the shale needed to be accounted for and
the correct hydrocarbon densities had to be estimated/calculated. For details o f the
porosity, statistical and spectral analyses refer to section 4.3 where explanations and
examples are provided.
Porosity histograms and statistical comparisons with core porosity values were
used to determine the validity o f the hypothesis that log porosity was overestimated in
horizontal wells. The results o f the analysis using a number o f porosity algorithms are
presented below. For each well there are two porosity histograms (0-40pu) for the
whole reservoir zone with two tables. In both porosity histograms the well 1 core
porosity values were scaled up proportionally by the ratio o f the number o f samples.
Note: well lz was compared with well lz core data over the oil zone, but well 1 core
data was used for comparison over the gas zone. The fluid parameters for the density
and acoustic shale corrected porosities were investigated in two ways. Firstly, using the
Oil Company’s (OC) fluid parameter values and secondly with values such that the
porosity equations are in agreement with well 1 core porosity values.
The (first) standard histogram includes the core, density, neutron, acoustic and
the Oil Company density and acoustic porosity values. The (second) calculated
histogram includes the core, square root, Wiley & Patchett, Gaymard & Poupon neutron
and Gaymard & Poupon density porosity values. The acronyms used in the histograms
are listed in Table 4.32.

XoneW i'll

xl 547-2270
Transition

Table 4.31 Top depths o f fluid zones for each well in metres.

4-82
 Chapter 4: Analysis o f the Data Sets

CPOR Well 1 core porosity

DenPor Shale corrected density porosity
NphiJS Sandstone neutron porosity
Nphi_Lime Limestone neutron porosity
AcouPor Shale corrected acoustic porosity
OC_DenPor Oil Company shale corrected density porosity
OCAcouPor Oil Company shale corrected acoustic porosity
SqPor Square root porosity
WPPor Wiley & Patchett porosity
GPNeut Gaymard & Poupon neutron porosity
GPDen Gaymard & Poupon density porosity

Table 4.32 Acronyms used for data set 2 histograms.

For each well, the first table provides statistical test results. The Student’s t-tests
for clear cells and Chi2 tests for grey cells, where a tick/cross (■/IX ) represents
acceptance or rejection o f the hypothesis at the 95% confidence level that there is zero
difference between that porosity equation and well 1 core porosity values over the fluid
interval stated [Table 4.31]. The second table gives details o f the fluid parameters used
in the equations for each fluid zone. In the fluid parameter table DEN refers to using the
fluid density in the density column with the two Gaymard & Poupon equations. OC
refers to using the relevant Oil Company’s fluid parameter in that particular porosity
equation. The oil zone is defined as oil at irreducible water saturation. The transition
zone is defined as oil at greater than irreducible water saturation.
The autocorrelation plots below all show the autocorrelation function for the
appropriate log and two coloured bands representing the significance o f the
autocorrelation function [Figure 4.23, Figure 4.26, Figure 4.29, Figure 4.34 and Figure
4.37]. If the autocorrelation function is outside the yellow band or outside the blue and
yellow bands, then the autocorrelation function is significant at that wavelength at the
95% or 99% confidence level respectively. The autocorrelation analysis for both wells
is summarised below in Table 4.33 giving the significant wavelengths at the 99%
confidence level. The autocorrelation plot for each log is given a letter that is referenced
in the top o f each cell in Table 4.25. The letter is a reference to the letter in the figure
referenced at the top o f the column. For example, the first cell in Table 4.33 refers the
autocorrelation plot for the well 1 gamma ray log, which is in Figure 4.23 plot i).

4-83
 Chapter 4: Analysis o f the Data Sets

0
a)
0-4,10-19, a) a)
0-60,
20-21,22-29, 0-50, 0-45, 50-80,
70-100,
32-33, 39-43, 125-245 130-210
190-340
45-47
b)
0-70, 80-120,
140-160, 162-
245
d)
d)
0-9,12-20,
0-120, 125-
24-26, 28-29,
150, 190-195
43-48

h)
h)
j) 0-20,40-55,
0-200,
All 65-75,
220-245
140-180
i) i) j)
k)
0-205, 0-50, 0-28, 30-38,
All
220-245 100-210 48-51

j)
0-15, 65-80,
140-180
e)
0-15, 30-40,
65-70, 75, 85,
110-145, MS-
152, 180-210
b)
0-7, 8-9,20-
21, 25-30, 33-
35, 44-45

a)
a)
0-10, 13-14,
0-10, 29-30,
20-22, 24-25,
46-48
26-40

4-84
 Chapter 4: Analysis o f the Data Sets

b)
( om pressional 0-40, 55-95,

travel tunc 1 0 5 -140,170-

220, 280-340

c)
Stone lev travel 0-52, 55-95,

150 -1 9 0 ,2 4 0 -

290, 305-340

d)
c)
Neutron 0-130, e)
0-13, 30-39,
porosits 160-165, 240 0-28, 34-55
43-48 65-190
245

b) 0-15, 20-40,
Dcnsits
0 -2 1 ,3 1 -3 2 , 1 4 0 -1 5 0 ,1 5 2 - 29-35,
correction 75-76,
46-47 1 6 0 ,1 6 5 -1 8 0 , 41-43,

185-200

Caliper

(Density)

C aliper 1

d)
C aliper 2
0-22, 22-55

penetration

Pressure

Table 4.33 Autocorrelation analysis giving the significant wavelengths in meters at the
99% confidence level. When all frequencies are significant at the 99% confidence level,
All is entered in the cell.

4-85
 Chapter 4: Analysis o f the Data Sets

Weill

The statistical tests demonstrated that density and acoustic porosity values both
matched the core porosity values over each fluid zone, although the match was not good
when the comparison was made over all intervals [Table 4.34]. The sandstone neutron
porosity values only match core porosity values in the oil, transition and water zones.
The Oil Company’s (OC) density and acoustic porosity values only match core over the
oil and transition zones and water zone respectively. O f the more complex porosity
equations Wiley & Patchett and Gaymard & Poupon neutron performed the best
matching in the oil, transition and water zones.
The standard histogram shows that the OC’s fluid parameters [Table 4.35] led to
overestimation o f the core porosity [Figure 4.21] whilst the density, neutron and
acoustic porosity equations match the core porosity values above 19% porosity. The
core porosity values around 10% porosity are due to high shale content in the core
samples in certain intervals. However, the removal o f these values does not alter the
results presented. The calculated histogram illustrates that the Gaymard & Poupon
neutron porosity underestimated porosity while the square root, Wiley & Patchett and
Gaymard & Poupon density porosity overestimated porosity [Figure 4.22].
The acoustic and density values used to match with the core porosity values and
the OC’s values were quite different. The OC’s density value was derived from the
water density at hydrostatic pressure and the OC’s fluid travel time values was that o f
water. However, the shallow investigation depth (~2inches) and the use o f OBM to drill
the well explains the use o f 210psec/ft, the approximate fluid travel time for oil. The
fluid densities used reflect the change in fluid density within the formation. These
results show that selection o f fluid parameters is important if porosity is to be estimated
with any accuracy. Interestingly, the OC’s density and acoustic porosity curves have
double peaks indicating the presence o f the different zones [Figure 4.21].
The autocorrelation analysis o f the well 1 wireline logs indicates that the logs do
not appear to be affected by the borehole or drilling process i.e. no correlation between
the caliper autocorrelation function and that o f any o f the other logs [Table 4.25 and
Figure 4.17]. The character o f the caliper [Figure 4.17e)] bears little resemblance to any
o f the other logs. There are broad similarities between acoustic travel time a),
compressional travel time b), density 1 f), density 2 g) and density correction h)

4-86
 Chapter 4: Analysis o f the Data Sets

autocorrelation plots. This adds confidence that the independent measurements (density
and acoustic) are observing the same geological information, since one m ay reasonably
expect them to react to noise differently. The gamma ray i) and resistivity j), k), 1) plots
are strikingly comparable due to the effect o f shale on the logs.

• Both density and acoustic porosity values provided the most reliable estimates o f
core porosity values for w elll.

G aym ard G aym ard

Porosity Sandstone Square W iley & * OC OC
Density Acoustic & Poupon & Poupon
equation N eutron Root P atchett Density Acoustic
N eutron Density
✓ X ✓ ✓ X X X X X
✓ ✓ ✓ X s ✓ X ✓ X
✓ ✓ ✓ ✓ ✓ ✓ X ✓ X
✓ ✓ ✓ ✓ ✓ ✓ ✓ X ✓
X X X X X X X X X

Table 4.34 Well 1 Student’s t-tests (clear cells) and Chi2 test (grey cells) results against
well 1 core porosity values, fluid parameters in Table 4.35.

0.55 210 OC OC DEN DEN 1.01 189

0.85 205 OC OC DEN DEN 0.70 189
0.85 210 OC OC DEN DEN 1.01 189
OC OC OC OC DEN DEN 1.01 189

Table 4.35 Well 1 fluid parameters.

4-87
 Chapter 4: Analysis o f the Data Sets

Well 1: Standard Porosity Histogram

500
— CPOR (%)
— DenPor
— Nphi_S
r

1-------------

HO O
8 o 'o '
^ Ql
P ® o
111
nPor
DuPor
300

3 K W
o-
<D
200
FA
II

100 ■| J

%
10 20 30 40
Porosity (%)

Figure 4.21 Well 1 standard porosity histogram.

Well 1: Calculated Porosity Histogram

500
— CPOR (%)
— SqPor
— WPPor
--- O r I'18U1
— GPDen

g* 300
c
d>
3
S"
it 200

100

0 H
10 20 30 40
Porosity (%)

Figure 4.22 Well 1 calculated porosity histogram.

4-88
 Chapter 4: Analysis o f the Data Sets

A u to c o rre la tio n f o r W e ll 1 W ir e lin e A c o u stic T ra v e l T im e L o g

■ r-0.01 ■ rO.Ol r0.05 r-0.05 DT r(L)

0.75 -

0.25 -

-0.25
Distance (m)

a)
Autocorrelation for Well 1 Wireline Acoustic Compressions! Travel Time Log
■ r-0.01 ■ rO.Ol r0.05 r-0.05 DTCO r(L)

.0.75 -

0.5 -

0.25 -

o 10 20 Distance (m) 30 40

b)
Autocorrelation for Well 1 Wireline Neutron Porosity Log
■ r-0.01 ■ rO.Ol r0.05 r-0.05 NPH Ir(L)

0.75 -

O 0.5 -

-0.25
D is ta n c e (m )

c)

4-89
 Chapter 4: Analysis o f the Data Sets

A u to c o r re la tio n f o r W e ll 1 W ire lin e P h o to e le c tric L o g

■ r-0.01 ■ iO.OI rO.OS r-0.05 PEFr(L)

0.75 -

-0.25
D istance (m)

d)
A utocorrelation for W ell 1 W ireline C aliper Log
■ r-0.01 ■rO.Ol r-0.05 CALTr<L)

0.75 -

*.25 -

-0.25
D istance (m)

e)

Autocorrelation for W ell 1 W ireline D ensity Log 1

■ r-0.01 ■ rO.Ol r-0.05 RHOB r(L)

.0.75 -

0.25 -

0 10 20 D is ta n c e (m ) 30 40

4-90
 Chapter 4: Analysis o f the Data Sets

A u to c o rre la tio n f o r W e ll 1 W ire lin e D e n sity L o g 2

■ r-0.01 r0.05 r-0.05 RHOB2 r(L)

0.75 -

B
0
1
!
It 0.25
im
-
3
<

-0.25
D istance (m)

g)

A u tocorrelation fo r W e ll 1 W ire lin e D en sity C orrection L og

■ r-0.01 ■ rO.Ol r0.05 r-0.05 DRHOKL)

0.75 -

1.25 -

-0.25
D ista n ce (m )

h)
Autocorrelation for Well 1 W ireline Gamma Ray Log
■ r-0.01 r0.05 r-0.05 GRr<L)

0.75 ■

0.25 -

-0.25
D ista n ce (m )

i)

4-91
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 1 Wireline True Resistivity Log

■ r-0.01 ■ rO.Ol 10.05 r-0.05 Rtr<L)

■S 0.5 -

0.25 -

0 10 20 Distance (m) 30 40

j)
Autocorrelation for Well 1 W ireline Deep Inducion Log
■ r-0.01 ■ rO.Ol r-0.05 — ILD r(L)

,0.75 -

0.5 -

0.25 -

0 10 20 Distance (m) 30 40

k)
Autocorrelation Tor Well 1 Wireline Medium Inducion Log
■ r-0.01 ■ rO.Ol r-0.05 ILMr(L.)

0.75 -

E
O
| 0.3 -

8
1
0.23 -

0 10 20Distance (m) 30 40

1)

Figure 4.23 Well 1 autocorrelation analysis o f the wireline logs. Note: x-axis is scaled
from Distance 0 to 45m, y-axis is scaled from Autocorrelation r(L) 0 to 1 above the x-
axis. Values outside the yellow and blue bands are significant at the 0.05 and 0.01
confidence levels respectively.

4-92
 Chapter 4: Analysis o f the Data Sets

Well lz

The statistical tests demonstrated that LWD density porosity values matched
both the well 1 and well lz core porosity values over the gas, oil and the whole
reservoir [Table 4.36]. This was only possible by use o f unrealistic density values in the
oil zone and the whole reservoir [Table 4.37]. The limestone neutron porosity values
matched core porosity values in the oil zone. Note: limestone neutron porosity is
approximately 4% porosity lower than sandstone neutron porosity in sandstone. The
OC’s density and complex porosity equations porosity values did not match core over
any interval. The standard histogram showed that the OC’s density porosity
overestimated the core porosity [Figure 4.24] whilst the limestone neutron porosity
equation matched the core porosity values. The calculated histogram illustrated that all
the complex porosity equations overestimated porosity [Figure 4.25 ].
The 4% porosity needed to match the core and LWD neutron porosity was
assumed to be a result o f standoff, >1 fin ch es o f effective standoff was needed to move
the values by 4% porosity. At >1 fin ch es standoff, density values are unusable. If the
standoff was constant along the entire well, then ‘limestone’ neutron porosity values
may be valid since the match with well lz core porosity values is good in the oil zone.
Neutron porosity logs are also more robust than density logs in the case o f standoff due
to their relative depths o f investigation. However, it was difficult to for see a physical
situation in which the LWD tool parallel standoff could be as large as 1finches in such
a well. It was possible that the LWD porosity tools were running skew across the
wellbore increasing the effective standoff observed by these dual detector tools.
The autocorrelation analyses o f the well lz LWD logs indicated that the logs did
not appear to be effected by the borehole or drilling process i.e. no correlation between
the rate o f penetration autocorrelation function e) and any o f the other logs [Table 4.25
and Figure 4.26]. In fact, there was little similarity between any o f the autocorrelation
plots, hence the problematic porosity values were likely to be a result o f tool
orientation/position whilst logging.

• Unrealistic fluid parameters were required to match the core porosity values [Table
4.37], therefore no reliable log porosity values could be calculated for well lz.
• These results show that there is no substitute for good quality log data.

4-93
 Chapter 4: Analysis o f the Data Sets

Well 1z: Standard Porosity Histogram

6000 -i----------------------
— CPOR (%)
— DenPor
5000 - — Nphi.Lime
—-OC DenPor

4000

D 3000
O’
LL
2000

1000

0 10 20 30 40
Porosity (%)

Figure 4.24 Well lz standard porosity histogram with well lz core porosity values.

Well 1z: Calculated Porosity Histogram

5000
— CPOR (%)
— SqPor
— WPPor
4000 - — GPNeut
—-GPDen

3000

O’
it£ 2000

1000

0 10 20 30 40
Porosity (%)

Figure 4.25 Well lz calculated porosity histogram with well lz core porosity values.

4-94
 Chapter 4: Analysis o f the Data Sets

✓ X X X X X X
✓ ✓ X X X X X
✓ ✓ X X X X X

Table 4.36 Well lz Student’s t-tests (clear cells) and Chi2 test (grey cells) results
against well lz core porosity values, fluid parameters in Table 4.37.

0.5 OC OC DEN DEN E01

0.0 OC OC DEN DEN 0.70

Table 4.37 Well lz fluid parameters.

4-95
 Chapter 4: Analysis o f the Data Sets

A u to c o rre la tio n f o r W e ll l z L W D G a m m a R a y L o g 1

■ r-0.01 ■ rO.Ol rO.OS r-0.05 GR1 r(L)

0.75 -

0.5 -

■§0.25 -

50 100 150 200

-0.25 -

-0.5
D ista n ce (m )

a)
A u to co rrela tio n fo r W e ll l z L W D G am m a R a y L o g 2
■ r-0.01 ■ ifl.Ol r0.05 r-0.05

0.75 -

0.25 -

100 150

-0.25
Distance (m)

b)
A u tocorrelation fo r W e ll l z L W D P re ssu r e L og

■ r-0.01 BrO.Ol r0.05 r-0.05 DDSr(L)

0.75 -

0.5 -

00 .2 5 -

150

-0.25
D is ta n c e (m )

4-96
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well lz LWD Neutron Porosity Log

■ r-0.01 ■ rO.Ol 10.05 r-0 .0 5 NPHIr(L)

0.75 -

0.5 -

0.25 -

100 150 200 V

-0.25
D istance (m)

d)

Autocorrelation for W ell lz LW D Rate o f Penetration Log

■ r-0.01 ■ rO.Ol r0.05 r-0.05— RO Pr(L)

0.75 -

0.25 -

100 150 200

-0.25
D istance (m)

e)
Autocorrelation for W ell lz LW D D ensity Correction Log

■ r-0.01 ■ rO.Ol r0.05 r-0 .0 5 ------- QUALf<L)

0.75 -

0 0.25 -

200

Distance (m)

4-97
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for W ell lz LWD Density Log

■ r-0.01 ■ rO.Ol rO.03

0.75 -

i
i
1 0.5 -

a
0.25 -

0 50 io o D is ta n c e (m )i5 0 200

g)
Autocorrelation for W ell lz LWD True Resistivity Log
■ r-0.01 ■ rO.Ol r0.05 r-0.05 Rtr(L)

0.75 -

0.25

100 150 200

-0.25
Distance (m)

h)
Autocorrelation for W ell lz LW D Resistivity Log
■ r-0.01 BrO.Ol r-0.03 EWRr(L)

0.75 ■

'0.25 -

100 150 200

-0.25
Distance (m)

i)

Figure 4.26 Well lz autocorrelation analysis o f the LWD logs. Note: x-axis is scaled
from Distance 0 to 240m, y-axis is scaled from Autocorrelation r(L) 0 to 1 above the x-
axis. Values outside the yellow and blue bands are significant at the 0.05 and 0.01
confidence levels respectively.

4-98
 Chapter 4: Analysis o f the Data Sets

Well 3

The statistical tests demonstrated that only the square root and Wiley & Patchett
porosity values matched both the core porosity values over all the units [Table 4.38]. Of
the other porosity equations, only the OC’s density and Gaymard & Poupon density
porosity values matched core over the oil zone [Table 4.39]. The wireline neutron
porosity values matched core over the whole reservoir. The histogram showed that the
square root and Wiley & Patchett matched the core, whilst the other methods
overestimated the core porosity except the Gaymard & Poupon neutron which
underestimated core porosity [Figure 4.27 and Figure 4.28].
The results show that the only change o f fluid parameters required to match the
core and density porosity values was the oil zone fluid density [Table 4.39]. No reason
for this could be found from the drilling information. The heavy value o f fluid density
required in the oil zone was due to the large number o f doggers encountered in this well
providing outlying data points. It was possible that a cuttings bed was lifting the
porosity tools from the floor o f the borehole, although poor tool application to the
borehole would provide this effect.

• Despite the apparently poor logs, the square root and Wiley & Patchett methods
provided good and reliable porosity values.

These results demonstrated that porosity in poor well conditions could be

achieved. Although the importance o f good quality density values in horizontal wells
should not be ignored as this would have further improved the results.
The autocorrelation analysis o f the well 3 logs showed similarities between all
the logs below 20m lag indicating that the logs were affected by the borehole conditions
[Table 4.25 and Figure 4.29]. The majority o f the energy being in the first few metres o f
lag. The gamma ray a) and rate o f penetration b) plot were characteristically the same,
the shale content controlling the drilling speed.

4-99
 Chapter 4: Analysis o f the Data Sets

Well 3: Standard Porosity Histogram

4000
— C PO R (:%>
— DenPor
— Nphi_S
OC_De nPor
3000 i

>
o
c

fjI
§o* 2000 1

£
Li-

1000

J jX
10 20
Porosity (%)
30
---------

40

Figure 4.27 Well 3 standard porosity histogram.

Well 3: Calculated Porosity Histogram

4000
— CPOR (%)
— SqPor
— WPPor
— GPNeut
— GPDen

>
o
c
§o- 2000
£
IL

1000

10 20 30 40
Porosity (%)

Figure 4.28 Well 3 calculated porosity histogram.

4-100
 Chapter 4: Analysis o f the Data Sets

X X ✓ ✓ X X X

X X ✓ ✓ X ✓ ✓
X ✓ ✓ ✓ X X X

Table 4.38 Well 3 Student’s t-tests (clear cells) and Chi2 test (grey cells) results against
well 1 core porosity values, fluid parameters in Table 4.39.

OC OC OC' DEN D EN 1.01

1.10 OC OC DEN DEN 0.70

Table 4.39 Well 3 fluid parameters.

4-101
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 3 Wireline Gamma Ray Log

■ r-0.01 ■ ifl.01 10.05 r-0.05-------- SGRr(L)

0.75 -

0.5 -

©0.25 -

100

-0.25
D istance (m)

a)

Autocorrelation for W ell 3 W ireline P hotoelectric F actor L og

■ r-0.01 ■ rO.Ol r0.05 r-0.05 PEF r(L)

0.75 -

0 0.25 -

100

-0.25
D istance (m)

b)
Autocorrelation for W ell 3 W ireline N eutron Porosity Log
■ r-0.01 ■ rOOl r0.05 r-0.05 -------- NPHIr(L)

0.75 -

00 .2 5 -

-0.25
D istance (m)

C)

4-102
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 3 Wireline Rate of Penetration Log

■ r-0.01 ■ rO.Ol r-0.05 RO PSrfl,)

0.75

0.5 -

0.25 -

100
-0.25 -

- 0 .5
D istance (m)

d)
Autocorrelation for W ell 3 W ireline D ensity Log

■ i>0.01 ■ rO.Ol r0.05 r-0.05 RHOB r(L)

0.75 -

0.5 -

0.25 -

100

-0.25
Distance (m)

e)
Autocorrelation for W ell 3 W ireline D ensity C orrection Log
■ r-0.01 ■ rO.Ol r0.05 r-0.05 DRHOr(L)

0.75 -

0.5 -

00 .2 5 -

100
-0.25
Distance (m)

f)

4-103
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 3 Wireline Caliper Log

■ r-0.01 ■ rO.Ol r0.05 r-0 .0 5 CALIr(L)

0.75 -

00.25 -

I
100 150 200

-0.25
D istance (m)

g)

Autocorrelation for W ell 3 W ireline True R esistivty Log

■ r^O.Ol ■ rO.Ol rt).05 r-0.05 RtJShift r(L)

0.75 -

0.5 -

'0.25

100

-0.25
Distance (m)

h)
Autocorrelation for W ell 3 W ireline Deep Induction Log
■ r-0.01 ■ rO.Ol r-0.05 H D r(L)

0.75 -

;0.25 -

100 150 200

-0.25 ■

-0.5
D istance (m)

i)

4-104
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 3 Wireline Medium Induction Log

■ r-0.01 ■ rO.Ol r0.05 r-0.05 ILMr(L)

0.75

iS 05
«
££0.25

|
0
100

-0.25
D ista n ce (m )

j)

Figure 4.29 Well 3 autocorrelation analysis o f the wireline logs. Note: x-axis is scaled
from Distance 0 to 210m, y-axis is scaled from Autocorrelation r(L) 0 to 1 above the x-
axis. Values outside the yellow and blue bands are significant at the 0.05 and 0.01
confidence levels respectively.

Well 3: filtered data

Well 3 and 6 data were filtered using a 19 point (~2m) moving average filter
(10cm sampled data). The choice o f this filter was determined by wishing to retain as
much o f the geological information in the logs whilst suppressing the noise as resulting
from borehole damage. The filtered data were processed again to evaluate the effect o f
filtering upon the porosity estimation from the logs [log plot in the wallet at the back o f
the thesis].
The statistical tests demonstrated that the filtered density, square root and OC
density porosity values matched the core porosity values over only the gas zones and
the sandstone neutron porosity only over the whole reservoir [Table 4.40]. The
histograms showed that the core porosity was overestimated except for the Gaymard &
Poupon neutron that underestimated core porosity [Figure 4.30 and Figure 4.31].
However, an unrealistic fluid density was used [Table 4.41]. The filtering did not
improve the accuracy o f the results, but in fact, the opposite was true [compare Table
4.38 with Table 4.40] and again the large number o f doggers may have been the cause
o f the bias in the data.

4-105
 Chapter 4: Analysis o f the Data Sets

Well 3: Standard Porosity Histogram

4000
— C P O R (% )
— DenPor
— N p h i_ S
— OC DenPor
3000

S'
c
0>
3 2000
O’
2>

1000

0
0 10 30 40
Porosity (%)

Figure 4.30 Well 3 standard porosity histogram with filtered data.

Well 3: Calculated Porosity Histogram

4000
— C P O R (% )
— SqPor
— W PPor
— G PNeut
3000 - — U H Uen

S'
c
<D
3 2000
cr
S>

1000

10 20 30 40
Porosity (%)

Figure 4.31 Well 3 calculated porosity histogram with filtered data.

4-106
 Chapter 4: Analysis o f the Data Sets

✓ X ✓ X X X /
X X X X X X X
X ✓ X X X X X

Table 4.40 Well 3 with filtered data Student’s t-tests (clear cells) and Chi2 test (grey
cells) results against well 1 core porosity values, fluid parameters in Table 4.41.

oc oc oc DEN DEN E01

1.10 oc oc DEN DEN 0.70

Table 4.41 Well 3 filtered data fluid parameters.

4-107
 Chapter 4: Analysis o f the Data Sets

Well5

The statistical tests demonstrated that the wireline density porosity values
matched the core porosity values over the oil, transition and water zones [Table 4.42].
The sandstone neutron porosity values matched core porosity values over the oil and
water zones, the acoustic only over the oil zone. The OC’s density porosity values
matched core over the oil, water zones and the whole reservoir and OC’s acoustic only
over the whole reservoir. Gaymard & Poupon neutron and Gaymard & Poupon density
porosity values matched core only in the water zone. The square root and Wiley &
Patchett porosity equations were poor, but matched core in the gas zone and, in the oil
and water zones respectively. The histograms showed that the OC’s fluid parameters
[Table 4.43] led to overestimation o f the core porosity [Figure 4.32] whilst the density
and acoustic porosity equations were a closer match with the core porosity values. The
other porosity equations overestimated porosity [Figure 4.33].
The 169psec/ft fluid travel time suggested that the acoustic measurement was
affected by cycle slipping, with p-wave suppression as gas would increase the travel
time required. The fluid density used in the gas zone being greater than the oil and
transition zones is thought to be due to the particularly shaley nature o f the gas zone
although mud invasion was a possible cause. If the gamma ray values from well 1 are
compared with well 5, the well 5 gamma ray values are significantly greater (~20 API).
Increased shale would indicate possibly lower permeability which would lead to deeper
invasion and therefore the tool seeing less gas. However, the shallow investigation
depth (~2inches) o f the acoustic measurement and the use o f OBM to drill the well
explained the 210psec/ft acoustic fluid travel time, the approximate fluid travel time for
oil. The fluid densities used reflect the change in fluid density within the formation
observed by the density tool, however the fluid densities used in the oil and transition
zones appear to be unrealistically low. Typical in-situ oil densities for North Sea
Tertiary aged reservoirs are 0.65 to 0.75g/cm3, 0.70g/cm3 was used by the Oil
Company.
The autocorrelation plots o f the well 5 wireline logs indicated that the logs all
had the same basic appearance. The plots gradually taper to approximately 30m, then
flatten to 55m lag. It was concluded that there was no borehole effect in the spectra o f
the logs [Table 4.25 and Figure 4.34].

4-108
 Chapter 4: Analysis o f the Data Sets

Well 5: Standard Porosity Histogram

750
— W1 CPOR (%)
— DenPor
— Nphi_S
— AcouPor •
— OC_DenPor
500 _— OC A couPor
>
o
c
<L>
3
cr
<d

250

-------- ^ —

10 20 30 40
Porosity (%)

Figure 4.32 Well 5 standard porosity histogram.

Well 5: Calculated Porosity Histogram

750
— W1 CPOR (%)
— SqPor
— -WPPor
— GPNeut
— GPDen
500

c
d>
3
o-
2>
LL
250

10 20 30 40
Porosity (%)

Figure 4.33 Well 5 calculated porosity histogram.

4-109
 Chapter 4: Analysis o f the Data Sets

G aym ard G aym ard

Porosity Sandstone Square W iley &
Density Acoustic & Poupon & Poupon
equation N eutron Root Patchett Density Acoustic
N eutron Density

Transition

Table 4.42 Well 5 Student’s t-tests (clear cells) and Chi2 test (grey cells) results against
well 1 core porosity values, fluid parameters in Table 4.43.

0.90 169 OC OC DEN DEN 1.01 189

0.50 210 OC OC DEN DEN 0.70 189
0.65 210 OC OC DEN DEN 1.01 189
OC OC OC OC DEN DEN 1.01 189

Table 4.43 Well 5 fluid parameters.

4-110
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 5 Wireline Acoustic Travel Time Log

■ r-0.01 ■ rO.Ol r0.05 r-0.05 DT r(L)

0.75 -

0.5 -

©0.25 -

10.01 20.02 30.02

-0.25
D istan ce (m )

a)
A u tocorrelation fo r W ell 5 W irelin e C a lip er D en sity L og

■ r-0.01 ■ rO.Ol r0.05 r-0.05

0.75 -

0.5 -

00.25 -

10.01 20.02 30.02 40.03 50.04

-0.25
D ista n ce (m )

b)
A utocorrelation for W ell 5 W irelin e C a lip e r L og 1

■ r-0.01 ■rO.Ol r0.05 r-0.05 C A Llr(L)

.0.75 -

0.00 10.01 20.02 30.02 40.03 50.04

D istan ce (m )

c)

4-111
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 5 Wireline Caliper Log 2

■ r-0.01 BrO.Ol r0.05 r-0.05- CAL2r(L)

0.75 -

0.5 -

0 0 .2 5 -

10.01 20.02 40.03

-0.25
D ista n ce (m )

d)
A u tocorrelation fo r W e ll 5 W ire lin e N eu tron P o r o s ity L og

■ r-0.01 ■rO.Ol r0.05 r-o.o: NPHIrCL)

0.75 -

0.5 -

0 0 .2 5 -

10.01 20.02 30.02 .40,03 [0.04

-0.25
D ista n ce (m )

e)
A u tocorrelation fo r W ell 5 W ire lin e D e n sity L o g

■ r-0.01 ■rO.Ol r0.05 r-0.05 IRHOBr(L)

0.75

0.5

00.25

10.01 20.02 30.02

■0.25
D ista n ce (m )

4-112
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well S Wireline Density Correction Log________

■ r-0.01 ■rO.Ol r0.05 r-0.05--------DRHOr(L)

..0.75 -

0.23 -

10.01 20.02

-0.25
D istance (m )

g)

A u tocorrelation fo r W e ll 5 W ire lin e G am m a R ay L og

■ r-0.01 BrO.Ol rO.OS r-0.05*

0.75 ■

*3 0.25 -

0.00 10.01 20.02 30.02 40.03 50.04

D istance (m )

h)
A u tocorrelation fo r W ell 5 W ire lin e P h o to electric F a cto r L og

■ r-0.01 ■ rO.Ol rO.OS r-0.05 PEF r(L)

0.75 -

B
e
3*
| 0.5
!
^ 0.25 -

0.00 10.01 20.02 30.02 40.03 50.04

Distance (m)

i)

4-113
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 5 Wireline Array Induction 90inch Log

■ r-0.01 ■ rO.Ol r0.05 r-0 .0 5 AT90 r(L)

0.75 -

0.5 -

00.25 -

10.01 20.02 30.02 40.03 50.04

-0.25
Distance (m)

j)

Figure 4.34 Well 5 autocorrelation analysis o f the wireline logs. Note: x-axis is scaled
from Distance 0 to 55m, y-axis is scaled from Autocorrelation r(L) 0 to 1 above the x-
axis. Values outside the yellow and blue bands are significant at the 0.05 and 0.01
confidence levels respectively.

4-114
 Chapter 4: Analysis o f the Data Sets

Well 6

The Chi2 tests demonstrate that acoustic and O C ’s acoustic porosity values did
not match the core porosity values through the well [Table 4.44]. The standard
histogram showed that the fluid parameters [Table 4.45] were inadequate to match the
core porosity values due to the spread in the values [Figure 4.35]. These results show
that selection o f acoustic tool for porosity was not good, due to the noisy acoustic log.
The choice o f the acoustic tool was likely to be because o f operational concerns over
nuclear sources in the well and the relative unimportance o f porosity data in this well.
Often the acoustic data is used for correlation with other wells and perforation zones are
selected on this basis.

The autocorrelation analysis o f the well 6 logs showed similarities between all
the logs below 15m lag indicating that the logs were effected by the borehole conditions
[Table 4.25 and Figure 4.37]. The majority o f the energy being in the first few metres o f
lag. The gamma ray a) and Stoneley travel time c) plot were characteristically similar
below 200m lag demonstrating the influence o f lithology on Stoneley waves. The
resistivity plots e) and f) were flat due to the near constant resistivity through the well.

Table 4.44 Well 6 Chi2 test (grey cells) results against well 1 core porosity values, fluid
parameters in Table 4.45.

210

Table 4.45 Well 6 fluid parameters.

4-115
 Chapter 4: Analysis o f the Data Sets

Well 6: Standard Porosity Histogram

2000
— C P O R (% )
— AcouPor
O C AcouPor

1500

S'
c
S 1000
O’
2>
Li.

500

0 10 20 30 40
Porosity

Figure 4.35 Well 6 standard porosity histogram.

Well 6: Standard Porosity Histogram

2000
— C P O R (% )
— A couPor
O C AcouPor

1500

3 1000

500

0 10 20 30 40
Porosity (%)

Figure 4.36 Well 6 standard porosity histogram with filtered data.

4-116
 Chapter 4: Analysis o f the Data Sets

Well 6: filtered data

The Chi2 tests demonstrate that the filtered acoustic and O C’s acoustic porosity
values did not match the core porosity values through the well [Table 4.46], The
standard histogram showed that the fluid parameters [Table 4.47] were inadequate to
match the core porosity values due to the spread in the values [Figure 4.36]. These
results showed that filtering the data had not improved the porosity estimate, mainly
relating to poor acoustic log data. Improving the acoustic log would have required
extensive reprocessing o f the acoustic waveform data that was not available. The
filtered log plot can be found in the wallet at the back o f the thesis.

Table 4.46 Well 6 filtered data Chi test (grey cells) results against well 1 core porosity
values, fluid parameters in Table 4.47.

210 189

Table 4.47 Well 6 filtered data fluid parameters.

4-117
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 6 Wireline Gamma Ray Log

■ r-0.01 BrO.Ol r0.05 r-0.05 GR r(L)

0.75 -1

0.5 -

00.25 -

49.99 99.97 149.96 199.95 249.94 299.92

-0.25
Distance (m)

a)
Autocorrelation for Well 6 Wireline Acoustic Compressions! Travel Time Log

■ r-0.01 BrO.Ol r0.05 r-0.05 DTCOr(L)

0.75 -

00.25 -

49.99 99.9- 149.96 249.94 299.92

-0.25
Distance (m)

b)
Autocorrelation for Well 6 Wireline Acoustic Stoneley Travel Time Log

■ r-0.01 arO .O l r0.05 r-0.05 DTST r(L)

0.75 -

00.25 -

49.! 299.92

-0.25
Distance (m)

4-118
 Chapter 4: Analysis o f the Data Sets

Autocorrelation for Well 6 Wireline Tension Log

■ r-0.01 ■ rO.Ol rO.OS r-0.05 — TENS r(L)

0.75 -

|
| 0.5 -

•3 0.25 -

49.99 99.97 149.96 199.95 249.94 299.92

Distance (m)

d)
Autocorrelation for Well 6 Wireline Deep Induction Log
■ r-0.01 ■rO.Ol i0.05 r-0.05 ILDr(L)

0.75 -

0.25 -

49.99 99.97 149.96 199.95 249.94 299.92

-0.25
Distance (m)

e)
Autocorrelation for W ell 6 W ireline Medium Induction Log

I r-0.01 ■ rO.Ol 10.05-------r-0.05----- ILMr(L)

0.75 -

49.99 99.97 149.96 199.95 249.94 299.92

-0.25
Distance (m)

Figure 4.37 Well 6 autocorrelation analysis o f the wireline logs. Note: x-axis is scaled
from Distance 0 to 340m, y-axis is scaled from Autocorrelation r(L) 0 to 1 above the x-
axis. Values outside the yellow and blue bands are significant at the 0.05 and 0.01
confidence levels respectively.

4-119
 Chapter 4: Analysis o f the Data Sets

4.5 Mud Invasion

Estimations o f the depths o f mud invasion for the three logging runs (vertical
well wireline, LWD and horizontal well LWD) were required to validate the statements
made in the previous sections regarding porosity calculations and photoelectric factor
anomalies. By modifying Harris’s equation [Equation 3.5] to give depth o f symmetrical
invasion from borehole wall, dit assuming that no mudcake exists, at t=0,

Equation 4.1
\ U J j

where t = time passed, p mf - mud filtrate viscosity, Kf = formation permeability, n =

borehole radius, r/= distance to limit o f reservoir, Pb = borehole fluid pressure and P /=
formation fluid pressure, <j>= porosity and Sxo - water saturation o f invaded zone. The
values used for the calculations are given below in Table 4.48.

Pmf 64 cP
n 8.5 inches
rf 538560 inches (2 miles). Reservoir compartmentalised by sealing faults (Dorn et al. 1996).
Pb 6374 psi calculated from depth and mud weight
Pf 4980 psi average of the repeat formation tester pressure data
* From calculated porosity

Kf From porosity/permeability relationship using the horizontal permeability for the vertical well
and the vertical permeability for the horizontal well [Figure 4.10].
S*o Calculated from porosity using the Oil company a, m and n values.

Table 4.48 Values used to estimate the depth o f mud invasion during logging.

The vertical well wireline mud invasion profile was calculated using a constant
time after circulation o f 36 hours. The vertical well LWD mud invasion profile was
calculated using time after circulation o f 80 hours plus the rate o f penetration multiplied
by the distance behind the bit o f the density/neutron tool. Figure 4.38 shows a
schematic o f the hypothesised invasion profile. The horizontal well LWD mud invasion
profile was calculated using the rate o f penetration multiplied by the distance behind the

4-120
 Chapter 4: Analysis o f the Data Sets

bit o f the density/neutron tool. Note: This method o f mud invasion depth estimation
assumes no mudcake and therefore mud invasion is likely to be shallower than
indicated by these calculations, although relative comparisons are informative.
The results are presented in Figure 4.39 for the vertical well and Figure 4.40 for
the horizontal well and demonstrate that in the vertical well the depth o f invasion was
large for both wireline and LWD logging runs. Note: Some' depth discrepancies are
apparent between the logs largely due to poor LWD depth data. The vertical well mud
invasion was typically >40inches and similar for both runs [Figure 4.39]. The depth o f
invasion was far greater than the depths o f investigation o f either o f the porosity tools
[Table 4.49]. The resistivity tools demonstrated little separation suggesting that the
invasion was either very shallow or greater than their depths o f investigation. The
wireline deep induction resistivity measured slightly greater resistivity values than the
other resistivity tools in the vertical well because it was measuring a greater volume
[Table 4.49 and Table 4.1]. However, the wireline and LWD porosity tools were
affected by the presence o f gas. The LWD porosity logs being affected to a greater
extent than the wireline porosity logs due to re-invasion o f gas into the invaded
formation around the borehole. Gas is highly mobile and can travel miscibly through
the invaded mud. The LWD logs were affected more than the wireline logs because the
LWD logs were run some 2 days after the wireline tools.
The mud invasion in the horizontal well was typically 4-10inches [Figure 4.40].
This explained why the LWD resistivity logs were very similar in value, with low
invasion therefore little influence on the resistivity [Table 4.1]. The maximum density
measured higher densities (lower porosity) than in the vertical well possibly due to
better flushing o f the gas from the near borehole region. The mud invasion profile could
also explain why the neutron porosity values were much lower than in the vertical well
[Table 4.1]. The depth o f invasion being 4-10inches would effectively remove the gas
effect from the density due to flushing, but gas would still affect the neutron due to its*
lOinch depth o f investigation.
The photoelectric factor logs may be explained by analogy with the mud
invasion profiles [Figure 4.39 and Figure 4.40], since the wireline and LWD caliper
logs agree well. The barite fines invaded into the formation with the mud, but due to
their density travelled more slowly than the mud invasion front creating a second barite
rich front. This barite rich front being time dependant and the shallow depth o f
investigation o f the photoelectric factor measurement resulted in high initial

4-121
 Chapter 4: Analysis o f the Data Sets

concentrations o f the barite (high PEF) close to the borehole [Table 4.49]. The
concentration o f barite and therefore PE F values decreased as depth o f invasion
increased. Alternatively, a more likely cause was that the LWD caliper log was
incorrectly indicating an in-gauge borehole and thus disguising the presence o f a
mudcake during LWD logging. Although mudcake could have been scraped o ff whilst
drilling as the stabilisers covering the detector are intended to do leaving poor
formation contact as the cause.

►
10cm
M------------------------------►
25cm

Figure 4.38 Schematic o f the mud invasion and tool positions.

4-122
 Chapter 4: Analysis o f the Data Sets

Depth of Invasion (Inches)

0 50 100 150 200 250 300 350 400
0 j_______ i_______ i_______ i_______ i_______ i_______ i_______ i_______

LWD Depth of invasion

Wireline Depth of invasion

Figure 4.39 Vertical well invasion profiles for LWD and wireline logging times.

4-123
 Chapter 4: Analysis o f the Data Sets

Depth oflnvasion (Inches)

0 10 20 30 40 50 60 70

LWD Depjh of invasion

Figure 4.40 Horizontal well invasion profile for LWD logging times.

4-124
 Chapter 4: Analysis o f the Data Sets

Wireline deep induction 60 [152]

Wireline medium induction 40 [102]
LWD amplitude resistivity 48 [122]
LWD phase resistivity 24 [61]
Wireline and LWD density 4 [10]
Wireline and LWD neutron porosity 10 [25]
Wireline and LWD photoelectric factor K[l]

Table 4.49 Tool depths o f investigation.

4-125
 Chapter 4: Analysis o f the Data Sets

4.6 Summary
4.6.1 Data Set 1
Factors affecting porosity calculations in the vertical well

The porosity analyses showed that for the vertical well the wireline density
porosity values were most reliable for predicting formation porosity values and had the
highest correlation with the core porosity values. The wireline neutron porosity values
were not as accurate as the wireline density porosity values, but in combination in more
complex porosity algorithms, their calculated porosity values were in good agreement
with core. The wireline acoustic porosity values were possibly affected by gas through
the Rotliegend gas interval. Alternatively, preferential acoustic energy propagation
along low porosity dipping beds encountered from the alternating dune slip face/lee
sands reduced the recorded travel times. By changing the fluid parameter the LWD
porosity values were improved by changes in fluid parameters to be in good agreement
with core porosity values over all the intervals expect for the LWD neutron porosity in
the Rotliegend gas interval.

• Square root single layer model using the LWD average density and neutron porosity
values provides best LWD derived porosity values in the vertical well.

Accepted practice is that the LWD maximum density is preferred, as its density
measurement is more valid and reliable with the LWD tools used (Rider 1996). The
difference between LWD maximum and average density complex porosity values were
small though tests showed closer match with core using LWD average density values.
The vertical well was logged with LWD (LAD) 2 days AFTER the wireline
tools. LWD porosity values were greater than their respective wireline porosity values
due to gas remigration towards the borehole decreasing the hydrogen index and fluid
density of the formation. The fluid parameters used indicated the effect o f the gas.
There was increased borehole rugosity/enlargement and thus worse formation to
detector contact than the wireline run indicated by the LWD caliper values.
Note that the standard method for obtaining fluid parameters is by crossploting
the core against log porosity values [Figure 4.41a, b and c]. The values derived from the

4-126
 Chapter 4: Analysis o f the Data Sets

crossplots compare well with the porosity values derived from the single layer model.
Principally the two methods are in essence the same, one graphical, the other numerical.

Vertical Well Core Porosity v. Density

Core Porosity (mA3/mA3)
0 0.2 0.4 0.6 0.8 1
0 - J------------------------------------- 1------------------------------------- L

W IR E L IN E D E N S IT Y (G /C C )
L W D M A X IM U M D E N S IT Y (G /C C )
L W D A V E R A G E D E N S IT Y (G /C C )
0.5 -
— L in ear ( W IR E L IN E D E N S IT Y (G /C C ))
- L in ear (L W D M A X IM U M D E N S IT Y (G /C C ))
Linear (L W D A V E R A G E D E N S IT Y (G /C C ))
y = - 1 . 6 1 8 ^ 2.65
R2 = 0.48 I
y = -1.8825X + 2.65
£ R2 = 0.34 I
«1.5 y = -2.0071x-» 2.65
LU
O R2 = 0.366

r
C a lc u la ted -luid Densitii 3S ( g /c m ^ )
W ire lin e 1 .0 3 2
L W D Maxirr lum 0 .7 6 8
2.5
^ * 8 L W D A v e ra g e 0 .6 4 3

a)

Vertical Well Core Porosity v.

Acoustic Travel Time
Core Porosity (mA3/mA3)
0.2 0.4 0.6 0.8
W IR E L IN E A C O U S T IC T R A V E L T IM E (U S E C /F T )

— Linear ( W IR E L IN E A C O U S T IC T R A V E L T IM E
(U S E C /F T ))
40

E 80 X x*
y = 129.85X- 55
R2 = 0.78; 2

120

160
Calculate) i Fluid A c o l istic T r a v e l'rim e

W ire lin e 1 8 5 8 5 us/s ec

200 ___
i __ i i.....

b)

4-127
 Chapter 4: Analysis o f the Data Sets

Horizontal Well Core Porosity v. Density

Core Porosity (mA3/mA3)
0 0.2 0.4 0.6 0.8 1
o -r LWD MAXIMUM
DENSITY (G/CC)

LWD AVERAGE
0.5 - DENSITY (G/CC)

— Linear (LWD
MAXIMUM
8 1
0
DENSITY (G/CC)) y = -2.7817x + 2 65
— Linear (LWD R2 = -0.07 4
AVERAGE y = -1.3398x + 2 . 6 5 ^ " *
DENSITY (G/CC))
|1 .5
ui
o
x

Calculated F uid Densitie s (g/cnV^)

LWD Maximijm 1.310
2.5 LWD Average -0.132

c)

Figure 4.41 Standard method o f core porosity against porosity log crossplots for
calculation o f fluid parameters.

Factors affecting spectral analysis in the vertical w ell

The autocorrelation analyses o f the vertical well logs indicated that the logs
appeared to be largely unaffected by the borehole or drilling process. The only
exception was the wireline PE F log that correlated with the LWD caliper log, which
displayed resonance features. The cause may have been that mudcake filled the spirals,
tool wobble or preferential mudcake build up on the more porous facies (dune slip
faces). The LWD density correction log was also affected and explained why the LWD
PEF and density logs were not similarly affected. The LWD density correction had

removed these features from the LWD PE F and density logs.

The vertical well wireline and LWD data was not filtered because this resulted in
only marginal improvement in the porosity calculations.

4-128
 Chapter 4: Analysis o f the Data Sets

Factors affecting porosity calculations in the horizontal well

LWD average density values were very low and LWD caliper values were large
which suggested that the borehole was enlarged and rugose. The values were large
enough to mean that both measurements were unreliable for any quantitative use and
cast doubts on the reliability o f all the other horizontal well LWD measurements.
Although it was possible using the LWD average density values to achieve good
agreement with core over some intervals the fluid densities used required for agreement
were unrealistic.
The absence o f core porosity data complicated the determination o f which
method(s) were appropriate to calculate accurate horizontal well porosity values.
Comparison of LWD log calculated porosity with the scaled vertical well core porosity
values over the relevant facies allowed the rejection o f methods that were inaccurate.

• The standard Wiley & Patchett LWD maximum density and neutron porosity values
provided the best LWD derived porosity values in the horizontal well.

This was not the same model as the vertical well (standard square root model)
suggesting that reliable and consistent porosity calculation from these LWD porosity
tools was not viable in this horizontal well and possibly all horizontal wells in this field.
The horizontal well LWD neutron and maximum density porosity values were
lower than their respective measurements in the vertical well. The horizontal well LWD
measurements were recorded approximately 18 hours after penetration and the vertical
well wireline measurements 36 hours were after penetration. The volume o f mud
invasion was similar in both cases when compared with the volume o f investigation o f
the porosity tools. The horizontal well LWD neutron porosity measurements were
recorded approximately 84 hours after penetration and therefore mud invasion was
more extensive. This was possibly due to barite mud invasion and mudcake increasing
concentration o f barite close to the borehole walls in the formation and mudcake at the
time o f logging compared with the vertical well. This would effectively have increased
the formation fluid density to greater than lg/cm 3 reducing the calculated porosity
values. The increased effect o f the barite is indicated by the very large LWD PEF
values in the horizontal well. The PEF measurement is only sensitive to the first Vi inch

4-129
 Chapter 4: Analysis o f the Data Sets

in front o f the short spaced detector indicating that the barite concentration is much
greater than in vertical well measurements.
Incorrect LWD tool position within the borehole whilst logging i.e. “riding the
borehole” leads to the possibility o f non-parallel standoff. This non-parallel standoff
can lead to either light or heavy bulk density values with no indication from the density
correction curve. Thus in a situation where the short spaced detector reads light (poor
contact) and the long spaced detector reads correctly (good contact) then the bulk
density would read too heavy. This situation is possible since the long spaced detector
leads the short spaced detector and the source down the borehole; the reverse o f a
wireline tool.

Factors affecting spectral analysis in the horizontal well

The autocorrelation analysis o f the horizontal well LWD logs indicated that
most of the logs were affected by the borehole or drilling process with only the LWD
resistivity log unaffected. The LWD density correction algorithm appeared to be unable
to correct sufficiently for the standoff encountered. The horizontal well LWD caliper
log demonstrated that spiralling was evident at a number o f wavelengths. The most
damaging wavelengths appeared to be the short wavelengths between 8-10 ft. The logs
were filtered accordingly to remove all wavelengths below 10 ft inclusively. The
porosity calculations were then repeated on the filtered horizontal well LWD data. The
additional knowledge o f the mud invasion was taken into account and fluid densities
greater than 1 g/cm3 were used.

Factors affecting filtered data porosity calculations in the horizontal well

The horizontal well filtered data LWD porosity values used were much
improved with respect core porosity values over most intervals.

• The best porosity values in the horizontal well were provided by the filtered LWD
maximum density and neutron porosity Wiley & Patchett single layer model.

4-130
 Chapter 4: Analysis o f the Data Sets

Factors affecting mud invasion in both wells

The vertical (1) and horizontal (lz) well demonstrated that in the vertical well
the depth of invasion was larger for both wireline and LWD logging runs. The vertical
well mud invasion was calculated to be >40 inches and similar for both runs. The
wireline and LWD porosity tools appear to be gas affected and the LWD porosity logs
being affected to a greater extent than the wireline porosity logs. This was because of
remigration of gas into the mud invaded formation around the borehole. The LWD logs
were affected more than the wireline logs because the LWD logs were run some 2 days
after the wireline tools. This was despite the calculated invasion since it was unlikely
that the mud invasion prevented gas migration towards the borehole (Woodhouse et al.
1991).
The mud invasion in the horizontal well was typically 4-10 inches, which
explained why the LWD maximum density measured higher densities (lower porosity)
than in the vertical well. However, the mud invasion profile could not explain why the
neutron porosity values were much lower than in the vertical well. The depth of
invasion would have had to have been greater than, but close to 10 inches throughout
the whole horizontal well to explain this situation.
The LWD photoelectric factor logs were explained by barite fines invading into
the formation with the mud, but travelled more slowly than the mud due to their
density. The slow moving barite rich mud resulted in high initial concentrations of the
barite (high PEF) close to the borehole. The concentration of barite and therefore PEF
values decreased as depth of invasion increased. Alternatively, a more likely cause was
that the LWD caliper log was incorrectly indicating an in-gauge borehole and thus
disguising the presence of a mudcake during LWD logging.

4-131
 Chapter 4: Analysis o f the Data Sets

4.6.2 Data Set 2

Porosity estimation in the two vertical wells (1 and 5) was aided by the core
porosity data available. Improving the porosity estimate compared with standard fluid
parameters only required a linear shift by adjusting the fluid parameters for the density
and acoustic logs. The sandstone neutron porosity was affected by shale, but the shale
correction used was ineffective. The use of the more complex porosity equations
(square root and Wiley & Patchett) were shown to be useful for wells 1 and 3 and to a
lesser extent well 5. One possible reason was the shale content of the sandstones was
not adequately estimated by the gamma ray volume shale transformation; for example
non-radioactive clays were encountered in well 5. Another is that the complex porosity
equations were designed for clean sands and their use was inappropriate.
Autocorrelation analysis demonstrated that there was little perturbation to the logs from
the borehole conditions, since the conditions were good in well 1, lz and 5.

• Both density and acoustic porosity values provided the most reliable estimates of
core porosity values for well 1.

Horizontal well lz porosity estimation was only possible due to the core
porosity data in well lz. The LWD log quality was poor especially the density that used
alone the logs would not have been any use for porosity estimation. The standoff
estimated at >114” rendered the LWD density log invalid. Porosity estimation was only
possible because the core porosity data allowed in-situ calibration of the neutron
porosity log, the neutron tool having a greater depth of investigation than the density
tool. As a result of the poor porosity logs, the more complex porosity equations were
unsatisfactory as porosity estimators principally due to the poor bulk density data.
Autocorrelation analysis demonstrated that there was little perturbation to the logs from
the borehole conditions. This is not surprising if the tools were effectively stoodoff by
>1!4” initially, the log responses will be relatively insensitive to further small additional
perturbations in standoff. The effective standoff was probably due to the LWD tools
running skew across the borehole axis i.e. non-parallel standoff causing the large
effective standoff and poor data quality. Another alternative is that the stabilisers

4-132
 Chapter 4: Analysis o f the Data Sets

nearest the bit were down cutting into the rock, “keyseating” the borehole and resulting
in the large standoff observed by the LWD porosity tools.

• Unrealistic fluid parameters were required to match the core porosity values,
therefore no reliable log porosity values could be calculated for well lz.

Porosity estimation in the horizontal wells 3 and 6 was not supported by core
porosity data. Well 3 shale corrected density porosity was a poor match with well 1 core
porosity values. It was possible that a cuttings bed was lifting the porosity tools from
the floor of the borehole, although poor tool application to the borehole would also
create this effect. The presence of numerous doggers in well 3 compared with wells 1,
lz and 5 may have influenced the results. Filtering data from well 3 did not improve
porosity estimation - in fact the results were considerably worse possibly due to
inappropriate use of the filtering in this case.

• Despite the apparently poor logs, the square root and Wiley & Patchett methods
provided good and reliable porosity values.

Well 6 acoustic data was of poor quality (noisy) presumably due to the nature of
the pipe conveyed logging run used. The poor acoustic travel time values may result
from the tool being tilted (non-parallel) and/or the acoustic energy short cutting through
fast beds (doggers). Autocorrelation analysis has shown the effect of borehole
conditions on the logs. Filtering data from well 6 did not improve porosity estimation.

These results demonstrated that porosity in poor well conditions could be

achieved (Well 3) although good well conditions are obviously preferable where they
can be achieved. Although the importance of good quality density values in horizontal
wells should not be ignored as this would have further improved the results.

4-133
 Chapter 4: Analysis o f the Data Sets

4.7 Conclusions
4.7.1 General data quality criticisms

For data set 1, data from only one vertical well and its horizontal sidetrack well
was made available. This meant that checking data quality with other comparable wells
in this field was not possible. The data set 1 LWD data from the horizontal well was a
splice of eight runs, however the overlapping sections of data were not available to
evaluate repeatability. For data set 2, data from three horizontal wells were available
from the field analysed with LWD data from one horizontal well and wireline pipe
conveyed data in the other two wells run with different toolstrings.
Only the product logs were available for the wells analysed because the data
was provided “as is” by Service Companies to Oil Companies. Little public domain
information is accessible regarding the detailed processing techniques used for the
wireline and LWD porosity tools employed, as this information is proprietary. This
includes filtering techniques, some of which are merely cosmetic and can obscure poor
quality data. However, if the ‘raw’ measurements were available, for example count
rates, the potential for developing alternative processing techniques would have been
possible. Oil Companies do double check the environmental corrections applied to
porosity product logs. This does not involve working from the “raw” count rates, but
from a “base” log that may itself be affected by incorrect processing.
A direct comparison of porosity values between wells is not necessarily valid.
Ideally, core porosity data would have been available for all the wells for comparison
with the log data. There was substantial scope for altering the position o f the core
porosity values in depth with respect to the horizontal well porosity data and thereby
changing the match achieved. A number of different approaches were tested and the
best comprise was to evenly distribute the core porosity values in depth. However,
some immunity was achieved by binning the porosity values into lpu bins prior to Chi2
testing. Note that log porosity values are accepted to be reliable to ±lpu. The lack of
core porosity data was problematic for accurate determination of which method(s) were
appropriate to calculate accurate horizontal well porosity values. However, comparison
of log calculated porosity with scaled vertical well core porosity values over the
relevant units allowed rejection of methods that were certainly inaccurate. A decision
was made to use data set 1 vertical well and data set 2 well 1 core porosity data for the

4-134
 Chapter 4: Analysis o f the Data Sets

statistical comparisons. Data set 2 well lz used well lz core porosity values because the
large number of core samples permitted their use. In addition, for data set 2 there was
little reason to assume that porosity was laterally constant, since the sands were shaley
and turbiditic with several doggers. However, without additional core porosity data no
conclusive assessment of porosity homogeneity could be made.
A further complication in the horizontal wells was that the borehole and
formation conditions differ from the vertical wells. Also calculating porosity was
problematic because the horizontal wells were not logged with the same class of tools
(wireline/LWD) or the same suite of tools (different measurements).
Details of the time-depth processing and the actual time-depth data are also
critical for good LWD data quality control, but were not available. Obviously, data may
be placed incorrectly with depth, a situation that becomes worse as the errors build as
the size of the time-depth file increases and the manner in which the calculations are
performed. In processing time-depth data it is possible, depending on if the porosity
processing technique uses data combined from several depths, that the porosity
calculations could be altered by different time-depth processing techniques. It is also
possible to dramatically alter the interpretation of the data depending on the time-depth
data. Bed boundaries and fluid contacts may be incorrectly located and the depth-
derived logs will be simply incorrect. The time-depth processing will also have the
effect of altering the effective volumes of investigation of the porosity tools making
comparisons between LWD and wireline tools more complex and probably invalid.
Wireline data is often on-depth to better than 1ft in 1000ft. LWD data is seldom so well
depth aligned, the author is aware of LWD data off-depth by more than 20ft in 100ft
and highly variable within a given 100ft section. Although absolute depth was not
particularly important for the research presented here, the relative depths were
necessary for matching core, wireline and LWD data prior to statistical comparison. In
addition, pipe conveyed log data are well known to suffer from poor depth control due
to the difficulties in maintaining cable tension during logging. This can render the
tension log useless as a means of checking data quality.
The LWD service company operations summary was available for the data set 1
horizontal well, but other operational information was not accessible. The operations
summary was available for all the data set 2 wells. Although the Oil Companies in
question applied strict quality control procedures problems still occurred, as was the
case for the data set 1 horizontal well, data set 2 wells lz, 3 and 6. There is no substitute

4-135
 Chapter 4: Analysis o f the Data Sets

to talking directly with the logging engineers involved at the time. One quality control
technique is to compare histograms of the data through the reservoir section from all the
wells in the field. This enables rapid identification of inconsistent data, unfortunately
this information was not available for data set 1.

*£ *
Differences from previous work

The validity of the porosity calculations in both vertical and horizontal wells
was tested against core porosity values from the vertical wells using Students’ t and
Chi2 tests. This has not been presented in such a manner previously for horizontal well
comparisons allowing significance or otherwise of the porosity transforms to be
evaluated.
The approach used in this chapter using numerous different porosity transforms
was based on the approach detailed in Cowan and Wright (1997). This approach had
not been applied to data in any published article known to the author. The techniques
used for dealing with shale/clay were standard.
Tool orientation considerations have been more seriously considered here than
in published papers particularly Cuddy et al. (1994) and Bedford et al. (1997). The
author feels that full consideration of the effects of tool positioning has not been
rigorously examined within the public domain. One clear indication of the deficiencies
of the particular LWD density and neutron tool used here was the introduction of the
azimuthal LWD density and neutron tools.
With respect to data set 1, horizontal wells drilled in this field after those
analysed in this thesis confirmed that the main source of error was the LWD tool
“riding the borehole” by the use of azimuthal LWD tools. The use of azimuthal LWD
tools enabled the Oil Company to design an algorithm to search for the maximum bulk
density in a more rigorous manner using the previous generations of LWD porosity
tools. Simultaneous maximum bulk density and neutron porosity values were calculated
to ensure their depth alignment. However, the techniques used within this chapter are
useful for the examination of historical horizontal well LWD data.

4-136
 Chapter 4: Analysis o f the Data Sets

4.7.2 Porosity equations

The use of square root and Wiley & Patchett methods provided good porosity
estimation in the Data set 1 horizontal gas well with poor borehole conditions
(spiralled) using the LWD maximum density filtered data. The data set 1 horizontal
LWD average density measurements were found to be unreliable as was the data set 2
horizontal well lz LWD density measurement. These logs were not suitable for any
quantitative use and cast doubts on the reliability of all the other horizontal well LWD
measurements.

• The use of the LWD average density measurement for porosity calculation is
strongly discouraged.

Spiralling of wells in both data sets led to further degradation of the porosity
measurements. The use of numerical filtering and careful selection of the fluid
parameters enabled the final porosity calculations to be improved significantly for data
set 1. Data set 2 would require the raw measurements for filtering to be investigated as
the porosity should be calculated from filtered count rates (Betts et al. 1990) and travel
times.
The quality of wireline and LWD log data in horizontal wells was generally
poor compared with vertical well log data. Standoff, borehole condition and noise
generated by the conveyance of the tools was of greater magnitude. The result was poor
quality data input into porosity equations. Data set 2 well lz LWD density measurement
was found to be unreliable for any quantitative use and the neutron porosity was ~4pu
too high. This cast doubts on the reliability of the LWD horizontal well measurements.
Well 3 wireline logs were noisy, but considered valid for porosity estimation. Porosity
estimation was still problematic due to the abundance of doggers encountered. Without
further study of other wells in the field, it would be difficult to conclude whether the
doggers are localised or continuous across the field. The vertical well logs may not be
resolving all the doggers whereas the relative angle of the beds to the horizontal wells
may be increasing the likelihood of observing the doggers. Well 6 acoustic porosity
estimation was poor. The spread in the acoustic travel time values demonstrated that the

4-137
 Chapter 4: Analysis o f the Data Sets

log was too noisy to be reliable for porosity estimation in this horizontal well due to
downhole noise from drill pipe and tool movement.
Filtering did not significantly improve the porosity estimation of wells 3 and 6
from data set 2. Although, the effect of spiralling of the well was thought to be a
secondary effect compared with incorrect shale volume estimation and poor tool
positioning within the borehole. The use of filtering may well be fruitful once shale and
standoff has been sufficiently accounted for.
For data set 2, the use of Gaymard & Poupon density and neutron methods were
not generally found to be useful for the horizontal wells, even using the filtered data.
These methods were designed for different situations to those found in the wells
examined above, due to the shale content. Accounting for shale using the gamma ray
log was insufficient for these methods. Neutron-density crossplot methods for shale
corrections were not considered as this has the potential of using the data for cross­
purposes and involve a degree of subjectivity. The shale corrected density porosity
provided the most reliable porosity in the wells considered in this chapter. However,
fluid densities were chosen to fit the expected porosity distribution rather than direct
use of vertical well fluid densities to predict porosity in horizontal wells.
Computer modelling could provide a family of Wiley & Patchett type equations
to meet the need for improved porosity and water saturation calculations in a broader
variety of mixed lithologies. Explicit use of lithology indicators such as gamma ray and
photoelectric factor log measurements and core mineralogy could enhance the quality
of porosity estimation, although good geochemical logging data would be preferable if
such a tool existed, as could the inclusion of invasion as suggested by Wiley & Patchett
(Patchett and Wiley 1994). At present the range of neutron-density porosity transforms
is limited to clean sandstone, limestone and dolomite although, accounting for shale in
calculations by the use of the volume shale calculations is often sufficient. There are
many cases where this is not so. The gap could be filled using Wiley & Patchett type
equations.

4-138
 Chapter 4: Analysis o f the Data Sets

Recommended porosity calculation practices

The following procedure is strongly recommended for porosity calculations for

horizontal wells:

• Check the logs for cyclicity (any repeated feature in the logs), especially the caliper,
photoelectric factor and porosity logs. If cyclicity is present, filter the logs before
porosity calculations are performed.
• The use of fluid densities, travel times and excavation constants from vertical wells
are likely to be incorrect for the horizontal well. Do try to match the porosity values
in the same interval/zone in the horizontal well as in the vertical wells, by adjusting
the fluid densities, travel times and excavation constants.
• The square root and Wiley & Patchett method provide good porosity values.
• Be aware that there are geological reasons for different porosity values in horizontal
wells that are in the same interval/zone as in the vertical wells. The closer the well
is physically to the vertical well(s) the more likely the porosity values are to be
similar.
• The use of the LWD average density measurement for porosity calculation is
strongly discouraged.
• Where shale content is an issue gamma ray logs need to be calibrated in known beds
to provide consistent and reliable volume shale estimation between wells and
differing tool types especially when wireline and LWD are used.
• Careful inspection of logs is required to identify what shifts are required by
justifying shifts from measured parameters e.g. standoff.

4-139
 Chapter 4: Analysis o f the Data Sets

Suggested Practical Remedies

• All raw measurements should be preserved for future use.

• Time-depth data should be preserved and rigorously checked for inconsistencies.
• If porosity, and therefore water saturation, are important in a horizontal well then
use wireline rather than LWD tools because the log quality is generally superior.
• The density tool must be included in the toolstring.
• Azimuthal LWD density and neutron tools should be used and run within the
correct rate of penetration range specified by the Service Company.
• LWD caliper measurement independent of the density measurements is essential.
This is now addressed by the introduction of an ultrasonic caliper.
• Sidewall core plug porosity would be a useful constraint on porosity derived from
logs run in horizontal wells.
• A shallow resistivity tool run in both wells would have provided evidence for
anisotropic mud invasion as a cause of the apparent porosity decreases.
• Improved clay/shale content information could be gained by the use of a reliable
geochemical logging tool or core mineralogy, especially when non-radioactive
clays are present.
• Drill smoother horizontal wells by the use of two stabilisers. One stabiliser within
say 10ft of the bit and a second one drill collar behind the first. The borehole may
become spiralled with only one stabiliser as a single stabiliser can act as a pivot.

Additional comments

• Down cutting of the first stabiliser could be forming a ledge enabling the LWD
tools to standoff leading to poor LWD density and possibly LWD neutron values.
• Ultrasonic caliper, wireline dipmeter or image logs in horizontal wells would verify
the state of the borehole. The improved borehole size values could be included in
the correction algorithms of the porosity tools to improve porosity estimation.
• Acoustic measurements in horizontal wells are likely to be noisy and unreliable for
porosity estimation or rock strength calculations.
• Redesign the porosity tools to improve tracking of the horizontal wells as tools are
often run in conditions for which they were not designed for.

4-140
 Chapter 5: Conclusions and Further Work

Chapter 5: Conclusions and Further Work

This chapter draws together the conclusions from previous chapters and is
divided into the following sections: recommended porosity calculation practices,
porosity equations, suggested practical remedies and further work.

Validation o f hypotheses

1. Log derived porosity values give poor estimates of formation porosity in horizontal
wells, regardless of tool type.
Data set 1 horizontal well (LWD) and data set 2 horizontal well 3 (wireline) log
porosity values both provided good agreement with the vertical well core
porosity values. However, data set 2 horizontal wells lz (LWD) and 6 (wireline)
log porosity values were not in agreement with the core porosity values as a
result o f poor quality log data.
2. LWD density-derived porosity provides the best LWD porosity tool estimate of the
true formation porosity in horizontal wells.
Data set 1 horizontal well LWD maximum density porosity values were in the
best agreement with the vertical well core porosity values. Though this can only
be shown to be case specific since only one well demonstrated this agreement.
3. Rugose and washed-out wellbore conditions in horizontal wells increases porosity
log values to a greater extent than in vertical wells.
The three vertical wells were all in good condition and so no conclusion can be
drawn on the relative effects o f the severity o f the welbore conditions between
vertical and horizontal wells.

The intention was to investigate all the hypotheses, but data was not available to
assess hypotheses 4, 5 and 6.

4. Formation fluid affects the magnitude of porosity anomalies in horizontal wells to a

greater extent than in vertical wells.
No shallow resistivity logs were available demonstrate the degree o f invasion.

5-1
 Chapter 5: Conclusions and Further Work

5. In horizontal gas wells, anisotropic mud invasion increases density derived porosity
values and decreases neutron porosity values, without increasing the apparent
variability in the measurements.
No azimuthal porosity and shallow azimuthal resistivity logs were available to
measure the effects o f anisotropic mud invasion on the recorded values.
6. Bedding and relative dip affect porosity calculations in wells.
No dipmeter, resistivity or acoustic imaging logs were available to measure the
dips and to evaluate the subsequent change or otherwise in response o f the
porosity values.

Differences from previous work

; The validity of the porosity calculations in both vertical and horizontal wells
were tested against core porosity values from the vertical wells using Students’ t and
Chi tests. This has not been presented in such a manner previously for horizontal well
comparisons allowing significance or otherwise of the porosity transforms to be
evaluated. The approach used in this thesis using numerous different porosity
transforms was based on the approach detailed in Cowan et al. (1997). This approach
had not been applied to well data in any published article known to the author.
Tool orientation considerations have been more seriously considered here than
in published papers particularly Cuddy et al. (1994) and Bedford et al. (1997). The
author feels that full consideration of the effects of tool positioning has not been
rigorously examined within the public domain. In data set 1, horizontal wells in this
field, drilled after those analysed in this thesis the Oil Company, confirmed that the
main source of error was the LWD tool “riding the borehole” by the use of azimuthal
LWD tools. The use of azimuthal LWD tools enabled the Oil Company to design an
algorithm to search for the maximum bulk density in a more rigorous manner using the
previous generations of LWD porosity tools. Simultaneous maximum bulk density and
neutron porosity values were calculated to ensure their depth alignment. However, the
techniques used within this thesis are useful for the examination of historical horizontal
well LWD data.

5-2
 Chapter 5: Conclusions and Further Work

5.1 Recommended porosity calculation practices

The following procedure is strongly recommended for porosity calculations in

horizontal wells:

• Check the logs visually for cyclicity, especially the caliper, photoelectric factor and
porosity logs. If cyclicity is present filter the logs from the raw measurements
before porosity calculations are performed.
• Use of the square root and Wiley & Patchett method provide good porosity values.
• The use of the LWD average density measurement for porosity calculation is
strongly discouraged.
• Where shale content is an issue gamma ray logs need to be calibrated in known beds
to provide consistent and reliable volume shale estimation between wells and
differing tool types especially when wireline and LWD are used.
• The use of fluid densities, travel times and excavation constants from vertical wells
are likely to be incorrect for the horizontal well. Do try to match the porosity values
in the same interval/zone in the horizontal well as in the vertical wells, by adjusting
the fluid densities, travel times and excavation constants.

5-3
 Chapter 5: Conclusions and Further Work

5.2 Porosity equations

Data quality

The quality of wireline and LWD log data in horizontal wells was generally
poor compared with vertical well log data. Standoff, borehole condition and noise
generated by the conveyance of the tools were often of greater magnitude. The result
was poor quality data input into porosity equations.
LWD logs are intended to record data within a few hours after penetration
avoiding poor borehole conditions and invasion effects [see Chapter 2]. LWD logs are
affected to a greater extent by perturbations than wireline logs are, possibly to the point
that the potential benefits of LWD log quality are often lost. Notably, the principal
porosity tool, density, is at a disadvantage in the LWD situation due to tool rotation and
the stiff tool length (20ft [6m] the effective stiff tool length could be much greater).
Wireline tools are able to conform to the borehole wall shape with only a lm pad and
caliper pushing the pad against the wall and reducing the effects of perturbations.
Direct comparison of porosity values between wells was not necessarily valid
since core porosity data (assumed to be true formation porosity) was not available for
all the wells for comparison with the log data. There was substantial scope for altering
the position of the scaled vertical well core porosity values in depth with respect to the
horizontal well porosity data and thereby changing the match achieved. The best
compromise was to evenly distribute the core porosity values in depth, though some
immunity was achieved by binning the porosity values into lpu bins prior to Chi2
testing. Note that log porosity values are accepted by the Oil Industry to be reliable to
±lpu. The lack of core porosity data was problematic for accurate determination of
which method(s) were appropriate to calculate accurate horizontal well porosity values.
However, comparison of log calculated porosity with scaled vertical well core porosity
values over the relevant units allowed rejection of methods that were certainly
inaccurate.

5-4
 Chapter 5: Conclusions and Further Work

Assessment o f porosity equations

• Shale corrected density porosity provided the most reliable porosity in the
dataset 2 vertical and horizontal wells.
• Wiley & Patchett method provided the most reliable porosity in the dataset 1
horizontal gas wells using the filtered LWD maximum density and neutron data.
• Gaymard & Poupon methods were not found to provide good porosity
estimation in horizontal wells, even when using filtered data.
• LWD average density was found to be unreliable for any quantitative use and its
use for porosity calculation is strongly discouraged.
• Spiralled horizontal wells resulted in the degraded porosity measurements.
• Filtering significantly improved porosity estimation for data set 1, but not data
set 2.
• Careful selection of fluid parameters enabled porosity calculations to be
improved for data set 1 and most wells for data set 2.
• Poor shale volume estimation and doggers significantly degraded the porosity
calculation for data set 2.
• Sonic logs are prone to downhole noise from drill pipe and tool movement in
horizontal wells.

Porosity equations

Data set 1 LWD average density logs were not suitable for any quantitative use
and cast doubts on the reliability of all the other horizontal well LWD measurements.
The data set 1 horizontal well LWD maximum density and neutron Wiley & Patchett
method were demonstrated to be useful for porosity estimation. The quality of wireline
and LWD log data in horizontal wells was generally poor compared with vertical well
log data. Standoff, borehole condition and noise generated by the conveyance of the
tools was of greater magnitude. The result was poor quality data input into porosity
equations.
Data set 2 well lz LWD density measurement was found to be unreliable for
any quantitative use and the neutron sandstone porosity was ~4pu too high. This cast

5-5
 Chapter 5: Conclusions and Further Work

doubts on the reliability of the LWD horizontal well measurements. Well 3 wireline
logs were noisy, but considered valid for porosity estimation. Porosity estimation was
still problematic due to the abundance of doggers encountered. Well 6 acoustic porosity
estimation was poor. The spread in the acoustic travel time values demonstrated that the
log was too noisy to be reliable for porosity estimation in this horizontal well due to
downhole noise from drill pipe and tool movement.
Filtering did not significantly improve the porosity estimation of well 3 and 6
from data set 2. Although, the effect of spiralling of the well was thought to be a
secondary effect compared with incorrect shale volume estimation and poor tool
positioning within the borehole. The use of filtering may well be fruitful once shale and
standoff has been sufficiently accounted for (requiring the raw measurements).
For data set 2, the use of Gaymard & Poupon density and neutron methods were
not generally found to be useful for the horizontal wells considered in this chapter even
using the filtered data. These methods were designed for different situations to those
found in the wells examined above due to the shale content. Accounting for shale using
the gamma ray log was insufficient for these methods. Neutron-density crossplot
methods for shale corrections were not considered as this has the potential for using the
data for cross-purposes and involves a degree of subjectivity. The shale corrected
density porosity provided the most reliable porosity in the wells considered in this
thesis.

5-6
 Chapter 5: Conclusions and Further Work

5.3 Suggested practical remedies

• All raw measurements should be preserved and readily available for future use.
• Time-depth data should be preserved and rigorously checked for inconsistencies.
• Filtering methods used by Service Companies should be made more transparent.
• If porosity and therefore water saturation are important in a horizontal well then use
wireline rather than LWD tools because the log quality is generally superior.
• A density tool must be included in the toolstring, LWD or wireline.
• Azimuthal LWD density and neutron tools should be used when using LWD tools
and run within the correct rate of penetration range specified by the Service
Company.
• Redesign the porosity tools to improve tracking in the horizontal wells.
• Ultrasonic caliper, dipmeter or image logs in horizontal wells would verify the state
of the borehole. The improved borehole size values could be included in the
correction algorithms of the porosity tools to improve porosity estimation.
• An LWD caliper measurement independent of the density measurements is
essential. This is now addressed by the introduction of an ultrasonic caliper
measurement.
• Sidewall core plug porosity would be a useful constraint on porosity derived from
logs run in horizontal wells.
• A shallow resistivity tool run in both wells would have provided evidence for
anisotropic mud invasion as a cause of porosity anomalies.
• Improved clay/shale content information could be gained by the use of a reliable
geochemical logging tool, especially when non-radioactive clays are present.
• Sonic measurements in horizontal wells are likely to be noisy and unreliable for
porosity estimation or rock strength calculations.
• Drill smoother horizontal wells by the addition of two stabilisers. One stabiliser
within 10ft of the bit and a second one drill collar behind the first. Inserting only
one stabiliser results in the borehole becoming more spiralled than without as a
single stabiliser will act as a pivot.

5-7
 Chapter 5: Conclusions and Further Work

5.4 Suggestions for further work

The scope for further work in horizontal well log analysis is near endless, some
techniques have only been touched on here. An important -note is that formation
properties can appear to change only very gradually or not at all in horizontal wells.
Often one of the most revealing techniques is to plot depth and on a compressed scale
(e.g. 1:1200/1:1000 normally 1:240/1:200 is used). Likewise log data on compressed
scales can draw attention to features that may otherwise be overlooked.
Four topics that deserve further investigation from the data presented above are:

• Computer modelling could provide a family of Wiley & Patchett type equations to
meet the need for improved porosity and water saturation calculations in a broader
variety of mixed lithologies. Explicit use of lithology indicators such as gamma ray
and photoelectric factor measurements could enhance the quality of porosity
estimation. As could the inclusion of invasion as suggested by Patchett and Wiley
(1994). At present the range of neutron-density porosity transforms is very limited,
(clean sandstone, limestone and dolomite), although shale is accounted for in
calculations by the use of the volume and is often sufficient. There are many cases
where this is not so. The gap could be filled using Wiley & Patchett type equations.
• Sonic measurements in horizontal wells. Dataset 2 well 6 sonic travel time values in
the sandstones were approximately lOpsec/ft slower than in the vertical wells.
Given this observation and evidence of slow sonic travel time values in the southern
North Sea (Austin et al. 1994), there appears to be further work in this line of study.
This is especially when the use of nuclear logs are unacceptable for operational
reasons, the sonic log may be the only porosity log available since NMR is not often
run in horizontal wells. Analysis of the full waveforms may shed light on the causes
of these anomalous horizontal sonic logs.
• Time lapse log analysis is possible with LWD and wireline tools and may provide
values of formation properties such as permeability.
• Zonation of log data. Concerns over the gamma ray for volume shale estimation and
the use of differing fluid zones raises the suggestion that the rather crude methods
employed above could be significantly refined. There is a substantial amount of

5-8
 Chapter 5: Conclusions and Further Work

published work regarding log zonation and automatic techniques for log evaluation.
However, no technique is yet able to cope with log data without user input.
Horizontal wells allow the subtle aspects of log evaluation to be put to the test. The
cut off between one ‘zone’ and another can be very diffuse.

Additional comments

Greater involvement with the Service and Oil Companies involved would have
provided more information regarding detailed porosity processing techniques. Increased
knowledge of the procedures and common practices involved in porosity and log
analyses of horizontal wells would have been beneficial. Planning of the data
acquisition needs to involve the Service and Oil Companies with specific aims set out
from the start. This includes the availability of “raw” measurements and processing
algorithms, and an agreement to share the results between all involved. Much of the
initial effort in this thesis involved the acquisition of data and the definition of the
project.

In addition the effect of trajectory and mud properties on horizontal log data
warrants closer examination. With subtle changes in formation properties in horizontal
wells changes to the borehole environment become ever more critical to log evaluation.
Resistivity tool response (induction and laterolog) in horizontal wells is a vast
topic in which the oil industry is only just starting to understand the problems involved.
New resistivity tools are being developed or in early field use with azimuthal
discrimination capablities. Computer modelling of horizontal well and 3D resistivity
responses is also in its early stages.
Azimuthal versions of the current logging tools are becoming available. The
data from such tools such as the LWD azimuthal density neutron, azimuthal laterolog
and 3D induction tools are aiding (horizontal) log interpretation and also uncovering
previously unknown log features.
A common theme throughout the push to understand horizontal well logs is
imaging. Not only the use of tools that provide 3D and 4D (time lapse) data, but
developments in log presentation and incorporation with other oilfield disciplines (core
analysis, seismics, well testing, reservoir modelling...) requires consideration for the
future of horizontal well log interpretation.

5-9
 Appendix 1: Geology of Data Set 1

Appendix 1: Geology of Data Set 1

The southern North Sea has been an area of active hydrocarbon exploration and
production since the early 1960s. This area covers what is known as the southern
Permian basin and stretches from the mid North Sea high in the west of the basin to the
Polish-Russian border in the east [Figure 1.1]. The hydrocarbon fields of the offshore
UK sector of the southern North Sea are predominantly gas producing [Figure 1.2].
The North Sea was formed as a result of tensional stresses in the Hercynian
foreland during the late Carboniferous/early Permian (Walker and Cooper 1987). At the
end of the Hercynian orogeny, a period of uplift effected the North Sea resulting in
erosion and accompanied by volcanism. The well data originate from a field in the Sole
Pit sub-basin which formed during the early Permian and lies off the North coast of
East Anglia, England [Figure 1.3].
The Sole Pit basin suffered inversion during the Late Jurassic/Early Cretaceous
that stripped away most of the Upper Jurassic sediments and again in the mid-Tertiary
as a result of the Alpine orogeny. The gas was derived from devolatalisied coal from the
underlying Carboniferous coal measures (Walker and Cooper 1987). The reservoir
rocks are predominantly Permian Rotliegend aeolian/fluvial sands. Generally, the
aeolian sand members have greater porosity than the fluvial members. Rotliegend
(Leman) sands were formed on the southern North sea basin margins. The primary
sediment source was erosion of the Carboniferous coal measures (Westphalian) from
the nearby London-Brabant massif that lay to the south [Figure 1.3](Ziegler 1975;
Glennie 1990). Ephemeral fluvial channels carried the sediment into the basin, and on
the flood plain the sediment was acted on by the strong trade winds from the North to
Northeast to form aeolian deposits. The Permian Zechstein mudstone (Kuperschiefer)
and the overlying evaporite sequences form the cap rocks.
Immediately overlying the Rotliegend is the basal mudstone of the Zechstein.
An initial rapid marine transgression led to an increase in moisture allowing a covering
of vegetation that later provided the basal mudstone of the Zechstein (Jenyon and
Cresswell 1987). Subsequently, the basin became extremely dry with intermittent
recharging of the basin that provided the source for the deposition of evaporites.

Al-1
 SyMtWKij

A l-2
Figure 1.1 Southern Permian basin (Plate 1 Brooks and Glennie 1987).
1
2? si
5 2888 !

raw OP COM“ 't4l!NTAL SLOPE OUTl :hlE OF ZFCMS-T Sl« UASt* SirrplitK d t's rr. if
ALPINE ' * *-OflWATIOK H »3«T ■" —- - f — — Sf tC C TIJC i FAULTS
Appendix 1: Geology o f Data Set 1

CtMSTLINt
 Appendix 1: Geology o f Data Set 1

I i HUTTC
jaw heather;
HEATHER'
fosiac*<r

MEIMD-L

IILLER

CLAYM
ORrptrT*"TAN
.
I ]'
IVARH0E*JR08 ROY.
[•MAUREEN
MABEL

'"El
{V1Y
albuskjE
NORE(Kll;
Josephine’

COUSL4

ESKOALEo

»L/2

^ ^UIDWAlf
eaarinc'TR-cauton
LiAELMAM.

:oevoroeV

9 OIL FIELO
C > GAS FIELD
HORNDEAF

NORTH SEA

200 km

Figure 1.2 Map o f North Sea oil and gas fields (Inside Cover o f Brooks and Glennie
1987).

Al-3
 Appendix 1: Geology o f Data Set 1

There are five main cycles of evaporites known, indicating at least five marine
transgressions during the Zechstein. Owing to the thickness of the evaporites (ignoring
later thickening due to halokinesis), the basin must have been regularly recharged with
saline water as indicated by numerous interbedded thin evaporites (Taylor 1990).
The data set used was from a 100-250ft(30-76m) thick Rotliegend gas reservoir.
In the vertical well, the 120ft(36.6m) Leman sands are underlain by the Westphalian
coal measures and overlain by a thin 2ft(0.6m) mudstone and dolomitic limestone
lft(0.3m) [Figure 1.4]. Above the dolomitic limestone were 146ft(44.5m) of anhydrite
with some thin dolomite interbeds. The anhydrite was capped by a minor gas reservoir
consisting of 29ft(8.8m) of dolomite. Above the dolomite, 8ft(2.4m) of anhydrite is
followed by a substantial thickness of halite/polyhalite.

Cleveland Anglo-Dutcli
Basin Basin

Cleaver Bank
High 54°N

East M idlands
Shelf

Figure 1.3 Field location map redrawn (Figure 1 Bulat and Stoker 1987).

A l-4
 Appendix 1: Geology o f Data Set 1

W vW I Halite

Anhydrite 8 ft (2m)

7~ 7~ 7,
Dolomite 30 ft ( I Om)

Anhydrite 40 ft (12m)
with dolomitic
interbeds

m : Dolomitic Limestone 1 ft (0.3m)

M udstone 2 ft (0.6m) |Zechstein, Kuperschiefer|

Sandstone 120ft (36.6m ) [Rotliegend, Leman Sands]

Carboniferous Muds. Silts and Sands

W m

Figure 1.4 Stratigraphic column based on the vertical well logs used.

NW Datum-Base Zechstein SE
Partially cemented zone of W eissliegend,,---------
t
o U ncem ented^^issliegend^------- ^ Interdune Dominated
o Sandstones \
Aeolian Dune and Interdune Sandstones-___ ^

oin
<N
Fluvial Sandstones

Basal Conglomerates
Carboniferous —^and Sandstones \

Figure 1.5 Schematic lithology cross section through the reservoir (Figure 2 Dom et al.
1996). Note: scale is only for guidance and is not necessarily representative.

The Rotliegend Leman sands are dominantly dune and interdune sands
overlying poorer quality fluvial sands (Dom et al. 1996). Thin cemented to partially
cemented Weissliegend overlies the Rotliegend dune sands. The reservoir thickens from
northwest to southeast with most of the thickening occurring in the low quality fluvial
sands [Figure 1.5]. The geological descriptions from core and/or cuttings from the well

Al-5
 Appendix 1: Geology o f Data Set 1

completion logs of the vertical and horizontal wells are given below [Table 1.1].
Reservoir quality varies rapidly laterally due to facies changes and
compartmentalisation resulting from fault related diagenesis (fault sealing)(Dom et al.
1996). The reservoir is transgressed by two northwest-southeast en-echelon fault zones
linked by a transfer zone consisting of discontinuous and variably oriented faults.
Several phases of reactivation have occurred leading to (possibly extensive) structural
controlled cementation (anhydrite/quartz). The wells have been positioned away from
these faults (potential flow barriers) as much as possible (identified using 3D seismics)
(Dom et al. 1996).

Sandstone: Poor to fair gas show, quartz sublitharenite becoming quartz arenite,
MD xl4-x236ft light brown to brown, black and white speckled, hard, fine to medium grained,
(232ft 70m) subrounded, well sorted, moderate to high sphericity, grain supported, lithic clasts,
finely laminated with finer grained laminae, trace heavy mineral grains, rare
TVD x811-x931ft chlorite near top and base, trace dark pyrobitumen stains near base, locally well
(120ft 36m) cemented by siliceous cement, poor visible porosity at top, good visible porosity at
base.

MD xl34-x906ft Sandstone: Poor to fair gas show, quartz sublitharenite to arenite, o ff white, light
(772ft 235m) grey, pale grey brown, fine to medium grained, moderately sorted, sub-angular to
subrounded, moderate sphericity, loose to friable, poor anhydritic or rare dolomitic
TVD x802-x835ft cement, anhydritic/argillaceous matrix, occasional red/black lithic clasts, poor to
(33ft 10m) fair visible porosity.
MD x906-x 1655ft Sandstone: Moderate to good gas show, quartz arenite, light grey brown, fine to
(749ft 228m) medium grained, moderately sorted, sub-angular to sub-rounded, moderate
sphericity, loose to friable, poor dolomitic or rarely siliceous cement, rare
TVD x835-x857ft argillaceous matrix, occasional red/black lithic clasts, fair to good visible porosity.
(22ft 6m)

MD xl655-x2060ft Sandstone: Poor to fair gas show, quartz sublitharenite to arenite, o ff white, light
(405ft 123m) grey, light grey brown, fine to medium, predominately fine grained, moderately
sorted, sub-angular to subrounded, moderate sphericity, loose to friable, poor
TVD x857-x846ft dolomitic cement, argillaceous or dolomitic matrix, rarely slightly anhydritic, rare
( l i f t 3m) red/black lithic clasts, poor to fair visible porosity.

Table 1.1 Well completion log geological descriptions based on Oil Company data.
MD = measured depth, TVD = true vertical depth.

Al-6
 Appendix 2: Geology of Data Set 2

Appendix 2: Geology of Data Set 2

The Palaeocene

The Palaeocene was initiated by the onset of spreading in the Labrador Sea
leading to the Maureen terrigenous sediment deposition in the northern North Sea
[Figure 2.2](Liu et al 1993). Subsequent uplift of the Scottish highlands and East
Shetland platform resulted with an uplift (effective sea level fall) of 375-525m,
followed by collapse of half that magnitude. Initial spreading of the Norwegian-
Greenland Sea and associated reorganisation of the Labrador Sea spreading influenced
the termination of the late Palaeocene sediment supply.
The Palaeocene sedimentation of the northern North sea was controlled by the
presence of three radiating basins of the Moray Firth, Viking Graben and Central
Graben [Figure 2.3](Mitchell et al 1986). These grabens were formed as a result of
volcanic updoming at their junction in Mid-Jurassic times. The early Palaeogene
sediments achieve a maximum thickness of 1200m over the Witch Ground Graben,
dipping and thinning in a wedge shape to less than 200m in the southern Central Graben
to the southeast (Stewart 1987). As the early Palaeogene sediments are traced to the
Inner Moray Firth, they thin and rise towards the Scottish coast.
Initiation of the Palaeocene saw the rejuvenation of faulting, inversion and
regional uplift of the sandstone rich Shetland and Scottish source areas [Figure
2.6](Galloway et al. 1993). An initial peak of the intrusive activity in western England
and Ireland occurred in the early Thanetian, coincident with the maximum sediment
supply during the deposition of the Andrew formation [Figure 2.4](Liu et al 1993).
Coincident with the deposition of the Andrew tuff was a sea level rise and followed by
the deposition of the Balmoral sequence [Figure 2.5].
The primary control on sedimentation was the progressive tilting and migration
of the uplift axis to the east with eustatic sea level changes playing an important
secondary control. The sedimentation of the first Palaeocene formation, Danian,
resulted in the reworking of the underlying Cretaceous Chalk (Ekofisk formation)
followed by the input of coarse siliciclasts through the Moray Firth (Maureen
formation) [Figure 2.7] (Galloway et al. 1993).

A2-1
 Figure 2.1
Field map with well locations (provided by the Oil Company).
Top Palaeocene Reservoir Well Locations

j | Gas cap
— — Initi GOC = !496 mTVDSS
Initial

|~~1 Oil leg

*9
Initial OWC = 554 mTVDSS
| | Aquifer

Fault

Appraisal well

Well trajectory

Appendix 2: Geology of Data Set 2

Horizontal oil producer

Gas injector
OWC monitoring
Proposed
 Appendix 2: Geology o f Data Set 2

. ? \ Brent >, '

J H l/fi
^
i East /
V ’1 ¥
'
I
/ Shetland f I FI'9 9 ^ '
/ Platform V / - f \ 9

H orst 0
F o rties

Figure 2.2 Sketch map of the northern North Sea (Figure 14.13 Mitchell et al 1986).

C ontinental e la s tic s

Paralic-deltaic

Shallow m arine

Volcanics

L a n d -n o
d ep o sitio n

SHETLAND . f h V °» °
-7 PLATFORM/* ’ V °'
'?i p<>ioo')!o°o
*'rt 1VQ.VV..
;7 //o^ tsira.

.SCOTLAND

ENGLAND/

Figure 2.3 Simplified palaeography of northern Britain and the North Sea in Middle
Jurassic times. The volcanics occur at the junction between the three radiating basins,
the Moray Firth basin (MFB), Viking Graben (VG) and Central Graben (CG).
Sediments are derived both longitudinally from the updomed triple junction and from
lateral, possibly fault-bounded, margins (Figure 14.15 Mitchell et al 1986).

A2-3
 Appendix 2: Geology o f Data Set 2

L I T H O S T R A T I G RAPHY Relative sea Expro Ages I

' Shelf level bio­ Periods
(Moray Grc
Slope->- Basin H igh________________ zones
Beauly Fm.

Alba
S eq u en ce
Alba Sand Formatioi

sequences

Beauly
Formation

T aySand ' Frigg/Tay

Formation S eq u en ce

ta u ly F o r m a tio n ^ 1
D o rn o c h F o rm a tio r Forties «
S eq u en ce
:orties Formati

Andrew Formation Andrew

S eq u en ce
Maureen Formation mreen Marl Fn

Chalk
S eq u en ce

Figure 2.4 Stratigraphic scheme for the Early Tertiary of the central North Sea/Moray
Firth area (Figure 2 Banner et al. 1992).

W ITC H G RO U N D GRABEN FO R T IE S CENTRAL

S O U T H HALIBUT TROU GH GRABEN

Figure 2.5 Early depositional sequences, central North Sea (Figure 2 Stewart 1987).
See Figure 2.7 for the definition of the numbers.

Continuation of siliclastic deposition followed during the Thanetian with a

progressive increase in source areas from the East Sheltand platform, but the Moray
Firth remained the focus of deposition. Three depositional episodes are known during
the Thanetian; lower Andrew, upper Andrew and Forties. Initially sand was deposited
onto the East Shetland platform from the highlands [Figure 2.8]. The coarse sands
formed steep sloped (3-5°) prograding-braided deltas that built out to the southeast. The

A2-4
 Appendix 2: Geology o f Data Set 2

sands were transferred into the basin from high relief via steep slopes in high viscosity
turbidite flows. Mudstones were deposited in the distal parts of the basin (Lista
formation). The end of the Palaeocene was a result of the Hebridean volcanism and the
opening of the Norwegian-Greenland sea causing the collapse of the Shetland platform.
This was marked by the deposition of Balder tuff.
Halokinesis in the northern North Sea was initiated during the Triassic and
Jurassic because of thermal subsidence and sediment accumulation (Hodgson et al.
1992). Salt walls and ridges were formed. The high sedimentation rate and reduction in
the supply of salt to the salt walls and ridges during the Cretaceous and early Tertiary
resulted in the salt ridges reforming into domes/diapirs spaced at approximately 8km
intervals along NE-SW Caledonian transfer fault systems and north-south Jurassic and
Permian fault trends.
During the Palaeocene, the salt diapir crests were condensed, but the flanks were
thickened leading to the development of peripheral rim-synclines. The result of these
salt movements was to effect the topography of the sea floor and therefore the
sedimentation during the Palaeocene. The Palaeocene turbidites were fed from the
Witch Ground Graben/Moray Firth basin (Hodgson et al. 1992). The salt related
topography provided the sediment with deposition sites on all flanks of the diapirs, but
the detailed deposition was controlled by the sediment influx direction, rate and the
position of the diapir.
The sea-floor topography dip was low, but the energy loss would be capable of
forcing high-density turbidite deposition. Another possibility is that the increased
turbulence when the flow hits a break of slope causes the flow to terminate [Figure 2.9].
Larger topography would cause the lower clastic-rich part of the turbidite flow to
deflect around the topographic high, while the muddier part continued over the top. The
energy imbalance caused by the flow separation resulting in the deposition of the clastic
part of the flow around the diapir [Figure 2.10]. Note that previous turbidite deposition
would also provide additional topographic features that would be capable of forcing the
same processes. The presence of the salt diapir topography and/or the continued
formation of the diapir as the sedimentation continued lead to the large four way dip
closed structures ideal for hydrocarbon accumulation e.g. a dome.

A2-5
 Appendix 2: Geology o f Data Set 2

UJ z

£
A

cE 4

II IW 7777AI ~ll \r777n\ ir I

NOSS36dWOO QW T3dOd 3Nld1W
7/////AY7777X\ ll~l I II
* ;
dfl W 3 8 8 V30NW 0 1 031V13H SW31SAS H D N 3 « * /1 3 I«

Z UJ
_l
v # j—

□
CO
“I

> z
UJ o

UJ I- o < <
3N 3 90 3N I1 3N 30O 31W 3 U 3M 01

Ayvuyai Sn030V13d0 O lS S V Id l
9, *

Figure 2.6 Synthetic palaeostress curve derived from the generalised sea-level curve for
the north-western European basins. Stresses are plotted relative to an arbitrary present
day stress level. Comparison with columns on the right-hand side with the timing of
tectonic events in north-western Europe shows correlation with tectonic phases in the
Atlantic, the north-western European rift system and the Alpine domain. Thick and thin
wavy lines denote major and minor rifting phases respectively. Major and minor folding
phases are indicated by thick and thin saw-tooth lines; stars denote phases of
anorogenic volcanism (Figure 4 Cloetingh et al. 1987).

A2-6
 Appendix 2: Geology o f Data Set 2

D£ PO SIT­
IONAL
WITCH GROUND GRABEN FORTIES CENTRAL
SEQUENCE
SOUTH VIKING GRABEN ' GRABEN

//V

FORTIES FAN

ANDREW FAN

MAUREEN FAN

SILSTONE MUDSTONE
ps51 HIGHSTANO m o s t
COAL
A PROOELTA MOST
CARBONACEOUS
SANDSTONE CALCAREOUS
(epi dot*-boring) MUDSTONE
SANOSTONE TUFFACEOUS
(«prdol*-fre«) TUFF

MELANGE/OEBRIS FLOW LIMESTONE

Figure 2.7 Composite chronostratigraphic diagram for the early Palaeogene of the
central North Sea (Figure 18 Stewart 1987).

A2-7
 Appendix 2: Geology o f Data Set 2

WITCH GROUND FORTIES

GRABEN FIELD
REGIONAL TILTING AND Rapid Regression
BASIN SUBSDENCE

HGHSTANO SYSTEM

Pronounced baslnward
shUt of coastai onlap. LOWSTAND SYSTEM

LOWSTAND FANS

HIGHSTAND SYSTEM
? hiatus

TRANSGRESSIVE SYSTEM

•hiatus-
LOWSTAND FAN
HIGHSTANO SYSTEM

TRANSGRESSIVE SYSTEM hiatus

TRAMP FED FANS----

LOWSTAND FAN
Rapid Regression
HIGHSTAND SYSTEM
— — hiatus - -
LOWSTAND FANS

BASIN SUBSDENCE
REJUVMATION OF -hiatus-
SHETLAND PLATFORM
PROVENANCE

50km 100km 150km

Figure 2.8 Coastal onlap curve for the early Palaeogene of the central North Sea. The
diagram shows the relative onlap position of each of the ten described depositional
sequences and depositional systems operating within each. The terms ‘highstand’ and
Towstand’ refer only to the relative position of sea level (Figure 19 Stewart 1987).

A2-8
 Appendix 2: Geology o f Data Set 2

SAND RICH SYSTEMS (Low efficiency)

FIow Direction

N
Rapid Turbulence
Palaeogene

Sand concentrated on upstream side of salt high

MUD RICH SYSTEMS (High efficiency)

M uds and Silts in

Turbulent Suspension Turbidity Current

Basal traction
carpet dominated Flow Stripping Salt Swell
by coarser grained
elastics

Sand deposition concentrated on leeward side of salt diapir

Figure 2.9 Turbidite-salt dome interaction; central North Sea (Figure 16 Hodgson et al.
1992).

A2-9
 Appendix 2: Geology o f Data Set 2

SALT TECTONICS/GEOMETRIES BASIN

EASTERN TROUGH CENTRAL GRABEN EVOLUTION

w
-1 0
Major period of clastic
-2 0 basin fill following therm al
uplift of Scotland -
-3 0 continued salt withdrawal
from synclines driving
-4 0
w idespread diapir growth.
-5 0

-6 0

-7 0

-8 0

-9 0 Creation of large salt

withdrawal synclines in
-100
major depocentres.
-110 Formation of salt ridges
on the margins of
-120 rem nant half-graben.

-130

-140

-150
Decoupling of basem ent
-160
and cover during late
Jurassic rifting. Inherited
-170 salt wall/sedim ent pod
geom etries reactivated.
-180

-190

■200
Salt withdrawal/dissolution
-210
wt creating accomodation
space forTriassic/M id
-220
Jurassic continental
•230 t
CL
sediments.
Z
■240

■250 rT T +V , Zechstein salt

deposition as post rift
•260
fill of Perm ian extensional
270 basins.lnitial period of
gravity creep.
-280

Figure 2.10 Salt controls on basin evolution; central North Sea (Figure 18 Hodgson et
al. 1992).

A2-10
 Appendix 2: Geology o f Data Set 2

The Andrew formation

The Andrew submarine fan is channellized with the channels persistent close to
the fan edges (Den Hartog Jager et al. 1993). The very high sedimentation rate reached
a maximum soon after the initiation of fan deposition. The sediment source was from
the northwest via the Fergus/Dornoch delta within the Moray Firth [Figure 2.11].
The channels lack cohesive levees due to the high sand volume of the turbidites
and therefore, the initial sand/shale distribution was erratic. However, as the sediment
supply waned, the levee and channel definition increased as the mud fraction of the
sediment increased [Figure 2.12]. This enabled the aggradation of the massive sand
units in the stable channels. The channels were up to 5m thick prior to amalgamation
forming a sheet-like fan. These massive channel sands are also interbedded with
mudstones from unconfined density flows. In addition, frequent local occurrences of
slump deposits are seen resulting from abrupt breakthrough of levees during channel
switching [Figure 2.13, Figure 2.14 and Figure 2.15].
The Andrew formation [58.5-59.5Ma] (Reynolds 1994) consists of amalgamated
fine to coarse grained mid fan lithic sands with mud clasts and carbonaceous fragments.
The beds within the formation show upward thickening cycles due to lobe switching.
The sediment is sourced from the East Shetland platform from three principal lobes
[Figure 2.16]. Note, the Andrew submarine fan sandstones have relatively constant high
acoustic velocity logs [Figure 2.16](Stewart 1987).
One particular example, the Andrew field [Figure 2.17], the Andrew submarine
fan sandstones are subdivided into a lower unit of thin bedded fine to medium grained
outer fan turbidites (Stewart 1987). The lower unit progrades into an upper unit of
medium and course grained channellized sandstones showing upward mid fan
thickening. The upper unit consists of thin-bedded fine-grained interlobe sandstones
with reworked sandy basaltic tuff at the top. Below are sands and muds with thin
reworked chalk and tuff layers (Knox et al. 1981). The lower unit comprises of massive
fine to medium grained poorly sorted sands with occasional coarse grains (Knox et al.
1981). The sands are relatively texturally uniform but with internal deformation
structures such as dish structures and at the shale contacts load structures. Some thin
beds show graded bedding and in the upper part cross lamination.

A2-11
 Appendix 2: Geology o f Data Set 2

ANDREW Delta plain—sand dominated

rw T30 4° e Positions at time of
Delta plain—mud dominated maximum progradation
!--- "~l Delta slope
1 I Thicker, sandier areas (N;G > 40%)
Slope-apron fans
$H1 Thinner, less sandy areas (N:G < 40%)
[: >■| Thicker, sandier areas (N:G > 40%)
Basin floor fans
] Thinner, sandier areas (N:G < 40-5%)
HI Muds (basin floor, ramp or shelf)

Sedim ent dispersal pattern

0 50 100 150 km
Divide between fans
Depositional/erosional edge 0 50 100 m ile s
(3; Numbers refer to discrete fans

Figure 2.11 Palaeogeographic map of the Andrew formation (Plate 1 Panel 2 Reynolds
1994).

ALBA
FORM ATION

FRIGG/TAY
FORMATION

RO G ALAN D
FORM ATION

FO RTIES
FORMATION

ANDREW
FORM ATION

Figure 2.12 Schematic changes in fan geometry through the Palaeocene and Eocene
(Figure 12 Den Hartog Jager et al. 1993).

A2-12
 Appendix 2: Geology o f Data Set 2

I n c i s e d V a lle y I n c is e d V a lle y
E x p o s e d S h e lf E x p o s e d S h e lf

In terflu v e

C h a n n e l-L e v e e
C anyon C anyon

H ig h s ta n d H ig h s ta n d
S y s te m s T ra c t S y s te m s T ra c t

A. Lowstand Fan L o w s ta n d F an
B. Early Lowstand Wedge L o w s ta n d F an

Channel-Levee Complex

In c is e d V aley< E x p o s e d S h e ll

P r o g r a d i n g C o m p le x

Late Lowstand W edge

(Prograding Complex)

H ig h s ta n d
S y s te m s T ra c t
Early Lowstand W edge
(C hannel-Levee Complex)
L o w s ta n d F a n
C. Late Lowstand Wedge D. Axial Section through
C h a n n e l-L e v e e
Prograding Complex C o m p le x Canyon and Slope

Figure 2.13 Block diagrams of the (A) lowstand fan, (B) early lowstand wedge
(channel-leeve complex), and (C) late lowstand wedge (prograding complex). An axial
section (D) through the canyon and onto the slope illustrates the stratal geometric
relationships between the three units. The lowstand fan is deposited where there is a
break in slope beyond the mouth of the canyon. The early lowstand wedge is deposited
within the canyon and downlaps onto the lowstand fan. The late lowstand wedge is also
deposited within the canyon and downlaps onto the early lowstand wedge (Figure 10.1
Posamentier and Erskine 1991).

A2-13
 Appendix 2: Geology o f Data Set 2

SECO NDARY
S ED IM EN T T R A N S P O R T
MAIN
CHANNEL

THIN BED D ED T U R B ID IT E S
A SSO C IA T E D W ITH C H A N N EL
M ARGINS A A BA N D O N M EN T
SEQ U EN CES

STA CK ED S E O U E N C E S
S A N D S T O N E S T U RBID ITES MUDDY D E B R IT E S

IN TERCH A N N EL LOW
ENERGY D ISTA L
T U R B ID IT E S A C L A Y ST O N E S

Figure 2.14 Forties formation submarine fan depositional model (Figure 6 Whyatt et al.
1992).

SAND

SEDIMENT TR A N SPO RT

Figure 2.15 Schematic diagrams for the depositional systems (Figure 6 Armentrout et
al. 1993).

A2-14
 Appendix 2: Geology o f Data Set 2

HAjLIBUT
H Q R S T^

: I T : O PPER PA N F V -
1 M - MIDDLE !FAN ^
- LiOWER i FAN i / 's.
T - lltlTER i i ___

SEQUENCE 3

Figure 2.16 Sequence 3 summary (Figure 8 Stewart 1987). Gamma ray logs are on the
left and sonic logs on the right-hand side of the logs shown.

In a second example, the Cyrus field [Figure 2.17], the Andrew formation
consists of a fining upward interbedded sequence passing into the Andrew mudstone
indicating the waning of the supply and fan abandonment (Mound 1991). The beds are
thin channel-like sandstones (0.5-2.5m) of limited lateral extent with interbedded
mudstones with a total thickness over reservoir of 10-15m thinning to the southeast and
thickening to the northwest. The upper unit porosities range from 12-18%, 50% water
saturation and permeability 50mD with 39-64% net/gross reservoir.
The lower unit is massive channellized mid-fan sandstone interspersed by thin
mud and limestones, but is vertically and laterally extensive throughout the reservoir
(Mound 1991). The lower unit has average porosities of 20%, water saturation 35% and
permeability 200mD with 90% net/gross reservoir based on the top 30m reservoir.

A2-15
 Appendix 2: Geology o f Data Set 2

v L IM IT O F S A N D S IN P A L E O G E N E
L IM IT O F M A T U R E U P P E R J U R A S S I C
J S O U R C E R O C K S (F R O M F IG .6 )
- E R O S IO N A L L IM IT O F T O P
PALEOCENE REFLECTOR

D E P T H T O T O P P A L E O C E N E (K m )

O IL F IE L D

G A S F IE L D

M A JO R FAULT
A X IS O F P A L E O G E N E
B A S IN
D IR E C T IO N O F
S E D IM E N T
TRA N SPO RT

I0MDAL

IL 0P N E R GST

GANNE'
Aberdeen

■JOANNE

lOOKm

Figure 2.17 Palaeocene play map. The shaded area shows the combined distribution of
the Paleogene and lower Eocene sandstones, mainly derived from the uplifted area of
the basin axis. Note that in many cases fields are not directly underlain by mature
source rocks (Figure 14 Pegrum et al. 1990).

A2-16
 Appendix 3: Excel Macro Listings

Appendix 3: Excel Macro Listings

nm nnm m m m m m m pon)sjty MaCIX) ... .

t f

' Macro 9/12/98 and the rest... by S Calvert '

' PhD Title "Log interpretation in Horizontal Wells" '

' at Leicester University '

11. This macro calculates porosity from an ascii log data file using a number o f '
' formulae. 1
' 2. Porosity is calculated using density and/or neutron tool porosity '
' data only. '
13. The porosity is calculated assuming a sandstone matrix in all cases. '
' 4. Solver is used iteratively to optimise the calculated porosity in some cases. '
15. Where code has not been written by myself, I have endevoured to reference the 1
' code to the relevent textbook or more often website. '

Option Explicit1Means all variables must be declared. Make finding error easier.
Option Compare T ex t' Means that text ooperations are case in-sensitive for SELECT CASE

1 S ta rt: Start of process time

' Finish: End o f process time
' Msg : Message to say it's finished and how long it took.

Dim i%, j%, k%, 1%, m%, s%, R%, a%, st%, sta%, stb%, stc%, std%, c%
Dim ResColumn%, NeutColumn%, DenColumn%, ResXoColumn%
Dim DepColumn%, CporColumn%
Dim NeutUnit, DepUnit, CporUnit As Boolean
Dim NeutStartRow%, NeutFinRow%, DenStartRow%, DenFinRow%, ResStartRow%, ResFinRow%
Dim ResXoStartRow%, ResXoFinRow%, CorStartRow%, CorFinRow%, FinRow%, StartRow%
Dim SwFinRow%, SwStartRow%, NumNeut%, NumDen%, NumRes%, NumResXo%, NumCalcPoints%, NumCalcSw%
Dim ws$, ps$, ips$, contents, msg$, mins%, secs!
Dim Finish!, Start!, Depth!, RhobM!, WeightM!
Dim Mud_Unit$, TempUnitS, Dep_Unit$
Dim Sal_Mud!, RW!, RwTemp!, SalWatTemp!, WatTempRhob!, ResPres!, ResTemp!
Dim ResRw!, RMF!, ResWatRhob!, ResSal_wat!, MaxCpor!, Numcor%
Dim NumSheets%, SourceDataSheetS
Dim DenVal(20000), WPNeutVal(20000), NeutVal(20000), ResVal(20000), ResValXo(20000) As Single
Dim Dep(20000), Cpor(20000), DenPor(20000), SqrtPor(20000), WPPor(20000) As Single
Dim ArchieSw(20000), NeutSw(20000), SqrtArchieSw(20000), WPArchieSw(20000) As Single
Dim GPPor(20000), GPArchieSw(20000), GPNPor(20000), GPNSw(20000), GPNWat(20000) As Single
Dim GPNeutSh(20000), GPDenSh(20000) As Single
Dim startcol%, histstop!
Dim ColDatal, ColData2, dr As Range

A3-1
 Appendix 3: Excel Macro Listings

Dim M eanl!, Mean2!, SD1!, nl% , n2%, SD2!, Correlation!, Critical!, Prob!, test!, cumul!
Dim p%, q%, plow!, phigh!
Dim comp$, allS, wks$
Dim Inital_Choice, Input_Data As Object
Dim Iter, Plot, Stat, ItPlot, ItPlot2, ItStat As Boolean
Dim NeutColFor!, CporColFor!
Dim Namelndex(lOO) As String
Dim histnum As Boolean
Dim ran, dan, ban As String

Const RhobH! = 0.5

Const WPCon! = 1.916
Const W PNeutl! = 0.4253
Const WPNeut2! = 0.001828
Const WPNeut3! = 0.00001885
Const WPDenl! = 0.4903
Const WPDen2! = 0.003882
Const WPDen3! = 0.004397
Const G PC onl! = 1.07
Const GPCon2! = 1.11
Const GPCon3! = 0.65
Const GPCon4! = 0.4
Const RhobSS! = 2.65
Const RhobW% = 1
Const RhobSS_W! = 1.65
Const CssRhobSS! = 2.650019

Sub Porosity()

' Use Dialog Boxes to obtain information.

11. Let the user set up the program with their choice o f calculations, stats and
' plotting.

Set Choose_Data_Sheet = New C h o o se D a taS h ee t

NumSheets = Sheets.count
For R = 1 To NumSheets
Sheets(R).Select
wks = ActiveSheet.Name
With Choose_Data_Sheet.Controls("OptionButton" & R)
.Visible = True
.Caption = wks
End With
Next

Load Choose_Data_Sheet
Choose_Data_Sheet.Show
For i = 1 To NumSheets
If Choose_Data_Sheet.Controls("OptionButton" & i).Value = True Then
ws = Worksheets(i).Name

A3-2
 Appendix 3: Excel Macro Listings

Else
Unload Choose_Data_Sheet
End If
Next
Unload Choose_Data_Sheet
Set Choose_Data_Sheet = Nothing

Set Initial_Choice = New Initial_Choice

Load InitialC hoice
Initial_Choice.Show
If Initial_Choice.Controls("Iter_Calc_Poro"). Value = True Then
Iter = True
Else
Iter = False
End If
If Initial_Choice.Controls("Calculate_Stats").Value = True Then
Stat = True
Else
Stat = False
End If
If Initial_Choice.Controls("Plot_Results"). Value = True Then
Plot = True
Else
Plot = False
End If
Unload Initial_Choice
Set Initial_Choice = Nothing

Set Input_Data = New Input_Data

Load InputD ata
With Input Data
.MudUnits.Addltem "(lb/gal)"
.MudUnits.Addltem "(g/cm3)"
.MudUnits.AddItem "(lb/cu ft)"
.MudUnits.Listlndex = 1
.DepthData. Value = 2600
.Depth_Unit.AddItem "(ft)"
.Depth_Unit.AddItem "(m)"
.Depth_Unit.ListIndex = 1
.Temp_Unit.AddItem "F"
.Temp_Unit.AddItem "C"
.Temp_Unit.ListIndex = 0
.Rw_Value. Value = 0.118
.RmfValue. Value = 10
.RwTemp.Value = 60
.BHT. Value = 216
.MudWeight. Value =1.415
.MudSalinity. Value = 144000
End With
Input_Data.Show

A3-3
 Appendix 3: Excel Macro Listings

RW = Input_Data.Rw_Value.Value
RMF = Input_Data.RmfValue. Value
RwTemp = Input_Data.RwTemp.Value
ResTemp = Input_Data.BHT. Value
TempUnit = Input_Data.Temp_Unit.Text
WeightM = Input_Data.MudWeight.Value
Mud_Unit = Input_Data.MudUnits.Text
Sal_Mud = Input_Data.MudSalinity .Value
Dep_Unit = Input_Data.Depth_Unit.Text
Depth = Input_Data.DepthData.Value
Unload Input_Data
Set Input_Data = Nothing

If TempUnit = "C" Then

RwTemp = 1.8 * RwTemp + 32
End If
If Mud_Unit = "(g/cm3)” Then
WeightM = WeightM / 8.345
Elself Mud_Unit = "(lb/cu ft)" Then
WeightM = WeightM / 7.48113
End If
If Dep_Unit = "(m)" Then
Depth = Depth * 0.3048
End If

' Insert a sheet for the calculations

Sheets.Add
ActiveSheet.Name = "Porosity"
ps = ActiveSheet.Name
Sheets(ws).Select

' Define name o f data sheet. Find the columns containing density and neutron data.

1Turn off screen updating to improve speed

Application.ScreenUpdating = False
Application.DisplayStatusBar = False

' Start the timer to work out how long this macro takes to run.

Start = Timer

F o r i= 1 To 30
contents = Cells(l, i)
Select Case contents
Case Is = "NPHI", "NPHI (%)", "NPHI (M3/M3)", "NPHI (PU)"
NeutColumn = i

A3-4
 Appendix 3: Excel Macro Listings

NeutColFor = Cells(50, NeutColumn).Value

Case Is = "DEPTH (FT)", "DEPT (FT)", "DEPT", "DEPTH", "DEPTH (M )”, "DEPT (M)"
DepColumn = i
Case Is = "RHOB", "RHOB (G/CM3)", "RHOB (G/CC)"
DenColumn = i
Case Is = "ILD", "ILD (OHM.M)", "RACM", "RACM (OHM.M)"
ResCoIumn = i
Case Is = "ILM", "ILM (OHM.M)", "RPCM", "RPCM (OHM.M)"
ResXoColumn = i
Case Is = "CPOR (PU)", "CPOR (%)", "CPOR (M3/M3)", "CPOR"
CporColumn = i
CporColFor = Cells(50, CporColumn).Value
End Select
Next i

If CporColumn = 0 Then
Iter = False
End If

' Find starts and ends o f density and neutron columns

NeutStartRow = FIRSTINCOLUMN(Cells(50, NeutColumn))

NeutFinRow = LASTINCOLUMN(Cells(50, NeutColumn))
DenStartRow = FIRSTINCOLUMN(Cells(50, DenColumn))
DenFinRow = LASTINCOLUMN(Cells(50, DenColumn))
ResStartRow = FIRSTINCOLUMN(Cells(50, ResCoIumn))
ResFinRow = LASTINCOLUMN(Cells(50, ResCoIumn))
ResXoStartRow = FIRSTINCOLUMN(Cells(50, ResXoColumn))
ResXoFinRow = LASTINCOLUMN(Cells(50, ResXoColumn))
If Not CporColumn = 0 Then
CorStartRow = FIRSTINCOLUMN(Cells(50, CporColumn))
CorFinRow = LASTINCOLUMN(Cells(50, CporColumn))
End If
%

' Find starts and ends o f columns and number o f points for calculations.
1Must have the same number o f points in the density and neutron columns,
' also for the resistivity column for Sw calculations.
t

If NeutStartRow > DenStartRow Then

StartRow = NeutStartRow
Else
StartRow = DenStartRow
End If

If NeutFinRow < DenFinRow Then

FinRow = NeutFinRow
Else
FinRow = DenFinRow
End If

A3-5
 Appendix 3: Excel Macro Listings

If StartRow > ResStartRow And StartRow > ResXoStartRow Then

SwStartRow = ResStartRow
Else
SwStartRow = StartRow
End If

If FinRow > ResFinRow And FinRow > ResXoFinRow Then

SwFinRow = ResFinRow
Else
SwFinRow = FinRow
End If

' Determine the format o f the Neutron and Core data

I

NeutUnit = NUMFORM(NeutColumn, StartRow, FinRow)

If Not CporColumn = 0 Then
CporUnit = NUMFORM(CporColumn, StartRow, FinRow)
End If

' Determine the number o f calculation points

NumNeut = NeutFinRow - NeutStartRow + 1

NumDen = DenFinRow - DenStartRow + 1
NumRes = ResFinRow - ResStartRow + 1
NumResXo = ResXoFinRow - ResXoStartRow + 1

If NumNeut > NumDen Then

NumCalcPoints = NumDen
Else
NumCalcPoints = NumNeut
End If

If NumCalcPoints > NumRes And NumCalcPoints > NumResXo Then

NumCalcSw = NumRes
Else
NumCalcSw = NumCalcPoints
End If

' Porosity Calculations '

ifftfiM m m m m tim m m m f tf fim m tm tiftm tm m itm m m m if ftftf m

' 1. Simple Density assuming sandstone matrix and water filled pores. '
12. Square root porosity using density porosity from 1. ((PHInA2+PHIdA2)/2)Al/2 '
' 3. Wiley and Pratchett Porosity using density porosity from 1. '
14. Gaymard and Poupon Hydrocarbon corrected Porosity using density porosity from 1.'
t I
m f f tff tm itiM m tm tm tttf tm m f tttff ttf im m tf fm m m m f tm tm fttftm

' Water Saturation Calculations '

A3-6
 Appendix 3: Excel Macro Listings

1. Use above porosity calculations assuming that a = l, m=n=2. '

2. Formation water salintiy is calculated from Rw at reservoir temperature and '
pressure. '

RhobM = WeightM * 0.12

Sal_Mud = Sal_Mud /1000000
SalWatTemp = ((3647.5 / (RW - 0.0123))A 1.04712) / 1000000
WatTempRhob = 1 + 0.7 * SalWatTemp
ResPres = Depth * (WeightM /19.2725)
ResRw = RW * (RwTemp + 6.77) / (ResTemp + 6.77)
ResWatRhob = (ResPres * 0.0000024337) + (SalWatTemp * 0.68) _
- (0.0000008 * ResTemp A 2) - (0.00006 * ResTemp) + 1.007
ResSal_wat = (ResWatRhob + (0.0000008 * ResTemp A 2) + (0.00006 * ResTemp) -1.007
- (ResPres * 0.0000024337)) / 0.68
MaxCpor = 0
Numcor = 0

F orj = 1 To NumCalcPoints
If Dep_Unit = "(m)" Then
Dep(j) = 3.28084 * Cells(j + StartRow - 1, DepColumn).Value
Else
Dep(j) = Cells(j + StartRow -1 , DepColumn).Value
End If
If Not CporColumn = 0 Then
If Not IsEmpty(Cells(j + StartRow -1 , CporColumn)) Then
If CporUnit = True Then
Cpor(j) = Cells(j + StartRow - 1, CporColumn).Value /1 0 0
Else
Cpor(j) = Cells(j + StartRow - 1 , CporColumn).Value
End If
If Cpor(j) > MaxCpor Then
MaxCpor = Cpor(j)
End If
Numcor = 1 + Numcor
End If
End If
DenVal(j) = Cells(j + StartRow -1 , DenColumn).Value
If NeutUnit = True Then
NeutVal(j) = Cells(j + StartRow -1 , NeutColumn).Value /1 0 0
Else
NeutVal(j) = Cells(j + StartRow -1 , NeutColumn).Value
End If
WPNeutVal(j) = NeutVal(j) * 100
ResVal(j) = Cells(j + SwStartRow -1 , ResCoIumn).Value
ResValXo(j) = Cells(j + SwStartRow - 1, ResXoColumn).Value
DenPor(j) = (RhobSS - DenVal(j)) / RhobSS_W
SqrtPor(j) = Sqr((NeutVal(j) A 2 + D enPor(j)A 2) / 2)

A3-7
 Appendix 3: Excel Macro Listings

WPPor(j) = (WPCon + (W PNeutl * WPNeutValG)) + (WPNeut2 * W PNeutVal(j)A 2)

- (WPNeut3 * WPNeutValG)A 3) + (W PDenl * (100 * DenPorG))) _
+ (WPDen2 * (100 * DenPorG))A 2) _
- (WPDen3 * (100 * DenPorG)) A 3 /100)) /1 0 0
ArchieSwG) = Sqr((ResRw / ResValG)) * (1 / DenPorG)A 2))
NeutSwG) = Sqr((ResRw / ResValG)) * (1 / NeutValG) A 2))
SqrtArchieSwG) = Sqr((ResRw / ResValG)) * (1 / SqrtPorG)A 2))
WPArchieSwG) = Sqr((ResRw / ResValG)) * (1 / W PPorG)A 2))
GPPorG) = (GPConl * DenPorG)) * (RhobSS - (1 - ArchieSwG)) _
* (GPCon2 + GPCon3 * Sal_Mud) - ArchieSwG) _
* (GPCon2 + GPCon3 * ResSal_wat)) / (RhobSS _
- (RhobM * (1 - ArchieSwG)) + (ResWatRhob * ArchieSwG))))
GPArchieSwG) = Sqr((ResRw / ResValG)) * (1 / GPPorG)A 2))
GPDenShG) = Sqr((RMF / ResValXoG) * DenPorG)A 2))
GPNWatG) = GPCon4 * NeutValG) * NeutSwG) * _
((Sal_Mud - ResSal_wat) / (1 - GPCon4 * Sal Mud))
GPNPorG) = NeutValG) + GPNWatG)
GPNSwG) = Sqr((ResRw / ResValG)) • (1 / GPNPorG)A 2))
GPNeutShG) = Sqr((RMF / ResVaDCoG) * GPNPorG)A 2))
Next j

If Abs(GPPorG) - GPNPorG)) > 0.05 Then

GPDenShG) = (1 / Sqr<(Rmf / ResVaDCoG)) * (1 < GPPorG)A 2))) _
- (GPArchieSwG) / Sqr((Rmf / ResVaDCoG)) * (1 / GPPorG)A 2))) - 1
GPPorQ) = (GPConl * DenPorG)) * (RhobSS - (1 - GPDenShG)) _
* (GPCon2 + GPCon3 * Sal_Mud) - GPDenShG) * RhobH _
* (1.33 - 0.036 * RhobH + 0.2 * RhobH A 2) - GPCon3 * ArchieSwG) _
* (Sal_Mud - ResSal_wat)) / (RhobSS _
' - (RhobM * (1 - GPDenShG) - ArchieSwG)) _
+ (RhobH * ArchieSwG)) + (ResWatRhob * ArchieSwG))))
GPArchieSwG) = Sqr((ResRw / ResValG)) * (1 / GPPorG)A 2))
GPNeutShG) = ( 1 1 Sqr((Rmf / ResValXoG)) * (1 / GPNPorG)A 2))) _
- (GPNSwG) / Sqr((Rmf / ResVaDCoG)) * (1 / GPNPorG)A 2))) -1
' GPNPorG) = NeutValG) + (NeutValG) * NeutShG) * RhobH _
* ((1.97 - 0.324 * RhobH + 1.8 * RhobH A 2 - 0.4 * Sal_Mud) _
/(1 - 0.4 * Sal_Mud)) + GPNWatG))
GPNSwG) = Sqr((ResRw / ResValG)) * 0 / GPNPorG)A 2))
End If

' Write calculations to Porosity worksheet

Sheets(ps).Select

Cells(l, 1) = "Depth"
If Not CporColumn = 0 Then
Cells(l, 2) = "Core Porosity"
End If
Cells(l, 3) = "Density Porosity"
Cells(l, 4) = "Neutron Porosity"
Cells(l, 5) = "Square Root Porosity"

A3-8
 Appendix 3: Excel Macro Listings

Cells(l, 6) = "Wiley&Prachett Porosity"

Cells(l, 7) = "Gaymard&Poupon Density Porosity"
Cells(l, 8) = "Gaymard&Poupon Neutron Porosity"
Cells(l, 10) = "Density Sw"
Cells(l, 11) = "Neutron Sw"
Cells(l, 12) = "Square Root Sw"
Cells(l, 13) = "Wiley&Prachett Sw"
Cells(l, 14) = "Gaymard&Poupon Density Sw"
Cells(l, 15) = "Gaymard&Poupon Neutron Sw"

Cells(l, l).Name = "Dep"

Nam elndex(l) = "Depth"
Cells(l, 2).Name = "CP"
Namelndex(2) = "CPor"
Cells(l, 3).Name = "DP"
Namelndex(3) = "DPor"
Cells(l, 4).Name = "NP"
Namelndex(4) = "NPor"
Cells(l, 5).Name = "SP"
Namelndex(5) = "SPor"
Cells(l, 6).Name = "WPP"
Namelndex(6) = "WPPor"
C ells(l, 7).Name = "GPDP"
Namelndex(7) = "GPDPor"
Cells(l, 8).Name = "GPNP"
Namelndex(8) = "GPNPor"

Cells(l, 10).Name = "SwDP"

Namelndex(lO) = "SwDPor"
Cells(l, 1 l).Name = "SwNP"
Namelndex(l 1) = "SwNPor"
Cells(l, 12).Name = "SwSP"
Namelndex(12) = "SwSPor"
C ells(l, 13).Name = "SwWPP"
Namelndex(13) = "Sw WPPor"
Cells(l, 14).Name = "SwGPDP"
Namelndex(14) = "SwGPDPor"
Cells(l, 15).Name = "SwGPNP"
Namelndex(15) = "SwGPNPor"

For k = 1 To NumCalcPoints
If Dep_Unit = "(m)" Then
Cells(l + k, 1) = Dep(k) / 3.28084
Else
Cells(l + k, 1) = Dep(k)
End If
If Not CporColumn = 0 Then
Cells(l + k, 2) = Cpor(k)
End If
Cells(l + k, 3) = DenPor(k)

A3-9
 Appendix 3: Excel Macro Listings

Cells(l + k, 4) = NeutVal(k)
Cells(l + k, 5) = SqrtPor(k)
Cells(l + k, 6) = WPPor(k)
Cells(l + k, 7) = GPPor(k)
Cells(l + k, 8) = GPNPor(k)
Cells(l + k, 10) = ArchieSw(k)
Cells(l + k, 11) = NeutSw(k)
Cells(l + k, 12) = SqrtArchieSw(k)
Cells(l + k, 13) = WPArchieSw(k)
Cells(l + k, 14) = GPArchieSw(k)
Cells(l + k, 15) = GPNSw(k)
N extk
Range(Cells(2, 1), Cells(FinRow, 15)).NumberFormat = "0.000"

If Iter = True Then

Call Iterative(Dep, Cpor, DenVal, NeutVal, WPNeutVal, ResVal, ResValXo, ResRw, _
GPDenSh, GPNeutSh, RhobSS, S alM u d , ResSal_wat, RhobM, ResWatRhob, NumCalcPoints,
StartRow, FinRow, Namelndex)
End If

' Calculate mean, standard deviation, correlation coefficients and Student T-tests

If Stat = True Then

If CporColumn = 0 Then
Set dr = Range(Cells(StartRow, 3), Cells(FinRow, 9))
histstop = Application.WorksheetFunction.Max(dr)
Else
histstop = MaxCpor
End If
histstop = 5 * (Int(histstop / 0.05) + 1 )
If histstop >= 100 Then
histstop = 100
End If
Sheets(ps). Select
SourceDataSheet = ps
st= 16
sta = 17
stb= 18
stc = 25
std = 26

10 Cells(l, sta) = "t-Tests assumes unequal variances for equal means at 0.05 significance"
Cells(5, sta).ColumnWidth = 20
If CporColumn = 0 Then
startcol = 3
Set ColDatal = Range(Cells(2, startcol), Cells(NumCalcPoints + 1, started))
nl = Application.WorksheetFunction.count(ColDatal)
Numcor = n l
Else
startcol = 2

A3-10
 Appendix 3: Excel Macro Listings

End If
cumul = 0
For c = 0 To histstop
If SourceDataSheet = ips Then
Cells(c + 2 0 ,35) = c * 0.01
Elself SourceDataSheet = ps Then
Cells(c + 20,27) = c * 0.01
End If
N extc
ran = Chr(64 + startcol) & StartRow & & Chr(64 + startcol) & FinRow
dan = Chr(66) & Chr(66) & 19
If SourceDataSheet = ips Then
ban = Chr(65) & Chr(64 + std - 26) & 20 & & Chr(65) & Chr(64 + std - 26) & (20 + histstop)
E lself SourceDataSheet = ps Then
ban = Chr(65) & Chr(65) & 20 & & Chr(65) & Chr(65) & (20 + histstop)
End If
Application.Run "ATPVBAEN.XLA!Histogram", ActiveSheet.Range(ran) _
, ActiveSheet.Range(dan), ActiveSheet.Range(ban), False, False _
, False, False
For c = 0 To histstop
If SourceDataSheet = ips Then
cumul = cumul + Cells(c + 20, 55) / Numcor
Cells(c + 20, std + 1) = cumul
E lself SourceDataSheet = ps Then
cumul = cumul + Cells(c + 20,55) / Numcor
If CporColumn = 0 Then
Cells(c + 20, std + 3) = cumul
Else
Cells(c + 20, std + 2) = cumul
End If
End If
Next c
With Range(Cells(19, 54), Cells((21 + histstop), 55))
.ClearContents
.Borders(xlEdgeTop).LineStyle = xlNone
.Borders(xlEdgeBottom).LineStyle = xlNone
.Borders(xlInsideHorizontal).LineStyle = xlNone
End With
For m = startcol To 7
Set ColDatal = Range(Cells(2, m), Cells(NumCalcPoints + 1, m))
Meanl = Application.WorksheetFunction.Average(ColDatal)
SD1 = Application.WorksheetFunction.STDEV(ColDatal)
nl = Application.WorksheetFunction.count(ColDatal)
If m = startcol Then
histnum = True
Else
histnum = False
End If
a = 9 * (m - 2)
Cells(2 + a, stb) = C ells(l, m)

A3-11
 Appendix 3: Excel Macro Listings

Cells(m, std) = C ells(l, m)

C ells(l, stc + m) = C ells(l, m)
Cells(m + 1, std) = C ells(l, m + 1)
C ells(l, std + m) = C ells(l, m + 1)
Cells(3 + a, stb) = M eanl
Cells(4 + a, stb) = SD1
Cells(3 + a, sta) = "Mean"
Cells(4 + a, sta) = "Standard Deviation"
Cells(5 + a, sta) = "Correlation Coefficient"
Cells(6 + a, sta) = "t Stat"
Cells(7 + a, sta) = "Critical t Stat"
Cells(8 + a, sta) = "Probablity in tail"
Cells(9 + a, sta) = "Comparison"
For 1 = m + 1 To 8
I f m = startcol Then
dan = Chr(66) & Chr(66) & 19
ran = Chr(64 +1) & StartRow & ":" & Chr(64 +1) & FinRow
End If
Set ColData2 = Range(Cells(StartRow, 1), Cells(FinRow, 1))
Call TtestStat(ColDatal, ColData2, M eanl, SD1, n l, n2, Mean2, SD2,
Correlation, Critical, Prob, test, comp, histnum, 1, ran, dan, ban)
s = st + 1
Cells(2 + a, s) = Cells(l, 1)
Cells(3 + a, s) = Mean2
Cells(4 + a, s) = SD2
Cells(5 + a, s) = Correlation
Cells(6 + a, s) = test
Cells(7 + a, s) = Critical
Cells(8 + a, s) = Prob
Cells(9 + a, s) = comp
Cells(m, s + 9) = comp
cumul = 0
I f m = startcol Then
For c = 0 To histstop
cumul = cumul + Cells(c + 20, 55) / n2
Cells(c + 20, s + 10) = cumul
N ex tc
End If
W ith Range(Cells(19, 54), Cells((21 + histstop), 55))
.ClearContents
.Borders(xlEdgeTop).LineStyle = xlNone
.Borders(xlEdgeBottom).LineStyle = xlNone
.Borders(xlInsideHorizontal).LineStyle = xlNone
End With
Next 1
N ext m
I f Iter = True Then
I f Iter = True And Stat = True Then
ItStat = True
Else

A3-12
 Appendix 3: Excel Macro Listings

ItStat = False
End If
ItPlot = True
Iter = False
Sheets(ips).Select
SourceDataSheet = ips
st = 24
sta = 25
stb = 26
stc = 34
std = 35
GoTo 10
End If
End If

Application.ScreenUpdating = False
Application.DisplayStatusBar = False

With Sheets(ps)
.Cells(19,27) = "Porosity"
.Cells(19,28) = "Core Porosity"
.Cells(19,29) = "Density Porosity"
.Cells(19,30) = "Neutron Porosity"
.Cells(19, 31) = "Square Root Porosity"
.Cells(19,32) = "Wiley & Pratchett Porosity"
.Cells(19,33) = "Gaymard & Poupon Density Porosity"
.Cells(19,34) = "Gaymard & Poupon Neutron Porosity"
End With

If CporColumn = 0 Then
Range(Cells(19,27), Cells((21 + histstop), 27)).Cut
Cells(19,28).Select
ActiveSheet.Paste
End If

If ItStat = True Then

With Sheets(ips)
.Cells(19,35) = "Porosity"
.Cells(19,36) = "Core Porosity"
.Cells(19,37) = "Density Porosity"
.Cells(19,38) = "Neutron Porosity"
.Cells(19, 39) = "Square Root Porosity"
.Cells(19,40) = "Wiley & Pratchett Porosity"
.Cells(19,41) = "Gaymard & Poupon Density Porosity"
.Cells(19,42) = "Gaymard & Poupon Neutron Porosity"
Range(Cells(19, 27), Cells(19, 34)).ClearContents
End With
End If

If Plot = True Then

A3-13
 Appendix 3: Excel Macro Listings

Call PlotGraph(ws, ps, ips, ItPlot, StartRow, FinRow, MaxCpor, Namelndex)

End If

Erase Cpor
I

1Turn on screen updating to view the mess

Application.ScreenUpdating = True
Application.DisplayStatusBar = True

' Work out time taken for job.

Finish = Timer
mins = Int((Finish - Start) / 60)
secs = Format((Finish - Start - mins * 60), "##0.0#")
msg = "Finished! Processing took " & mins & " minutes and " & secs & " seconds."
MsgBox (msg)

End Sub
I
im t m ttm m tm tt im m f tim m tm m m m m t m tm m f ti H m im tm t f t

1Function to calcalate num eric format o f a column '

itf tf ttttf ttiitttm ttiittH iitf tf ff itm iff iifititttftttif tm m iitf ff tittitm tt?

t t

Function NUMFORM(TestCol, StartRow, FinRow As Integer) '

Dim z%, num%, bill$, ben$ '
Dim fraction, fsum As Integer '
Application.Volatile '
num = StartRow 1
F o r z = 1 To 10 1
22 If Not IsEmpty(Cells(num, TestCol)) Then '
bill = Format(Cells(num, TestCol), "Scientific") '
Else ’
num = num + 1 '
GoTo 22 1
End If
ben = Right(bill, 3) '
If Left(ben, 1) = Then '
fraction = 0 '
Elself Left(ben, 1) = "+" A nd Left(Right(bill, 8), 4) = "1.00" _
Or Left(Right(bill, 8), 4) = "0.00" Then
fraction = 0 '
Else '
fraction = 1 '
End If
fsum = fsum + fraction '
num = num + 5 1
Next '
If fsum = 0 Then '
NUMFORM = F a lse ' 0 to 1

A3-14
 Appendix 3: Excel Macro Listings

Else '
NUMFORM = True ’ 0 to 100
End If
End Function '

m m t tm m t tm ttf ii iif tttt tm m ti itii iif im tm ttim tm itt m m tM i m tf i

i m t f t m t m f m im tm f tm m m m tm m t m ttm m tm m t m tm t m i

' Function from John W alker at http://www.j-walk.com/ss/excel/index.htm '

’ The function returns the last filled cell row number in a given column. '
' This probably the best Excel VBA web site. '

Function LASTINCOLUMN(mgInput As Range) '

Dim WorkRange As Range '
Dim i As Integer, CellCount As Integer '
Application.Volatile '
Set WorkRange = mgInput.Columns( 1).EntireColumn '
Set WorkRange = Intersect(WorkRange.Parent.UsedRange, WorkRange) '
CellCount = WorkRange.count 1
For i = CellCount To 1 Step -1 '
If Not IsEmpty(WorkRange(i)) Then '
LASTINCOLUMN = i
Exit Function '
End If
Next i ’
End Function '

m m m tiim tttm m m m m i m m tt ttm tm tm M tf tm m f f in t tttm t

fm m m m m tiM f ttm tH m m ttm ttm im m m m m tm tfttfM ttM itH f

' Function from adapted John Walker at http://www.j-walk.com/ss/excel/index.htm '

' The function returns the first filled cell row num ber in a given column, '
' ignoring the header. '
tttm tf tm tm tm tm m tttm ttm t M m tm t ttm t ttim i m tm f itt f i m m

Function FIRSTINCOLUMN(mgInput As Range) '

Dim WorkRange As Range '
Dim i As Integer, CellCount As Integer '
Application.Volatile '
Set WorkRange = mgInput.Columns( 1).EntireColumn '
Set WorkRange = Intersect(WorkRange.Parent.UsedRange, WorkRange) '
CellCount = WorkRange.count '
For i = 2 To CellCount '
If Not IsEmpty(WorkRange(i)) Then 1
FIRSTINCOLUMN = i
Exit Function ’
End If
Next i '
End Function '

A3-15
 Appendix 3: Excel Macro Listings

tim m tim ttitf tt tm f iiit tttim tm tt tm tm ttf tttf ttt m iit tm m tt m tf t

1 T Tests and Statistics 1

ttim tm im m m m im tim m m m m tim tm iim tim m tm im iim

i »

' The idea here is to calculate the mean, standard deviation o f the porosity '
' calculations and t test them against one another and the original data and 1
' also core data if available. '

Sub TtestStat(ColDatal, ColData2, M eanl, SD1, n l, n2, Mean2, SD2, _

Correlation, Critical, Prob, test, comp, histnum, 1, ran, dan, ban)

Dim SDDiff!, dffi, df%

Const HypDiff% = 0

Mean2 = Application. WorksheetFunction.Average(ColData2)

SD2 = Application.WorksheetFunction.STDEV(ColData2)
n2 = Application.WorksheetFunction.count(ColData2)
Correlation = Application.WorksheetFunction.Correl(ColDatal, ColData2)
SD D iff= (SD1 A 2 / n l) + (SD2 A 2 / n2)
dff = ((SD D iffA 2) / (((SD1 A 2 / n l ) A 2 / (n l -1 )) + ((SD2 A 2 / n 2 ) A 2 / (n2 -1))))
If d ff> = (Int(dff) + 0.5) Then
d f = Int(dff) + 1
Else
d f = Int(dff)
End If
Prob = Application.WorksheetFunction.TTest(ColDatal, C olD ata2,1, 3)
test = (Meanl - Mean2 - HypDiff) / Sqr(SDDiff)
Critical = Application.WorksheetFunction.TInv(0.1, df)
I f Abs(test) > Critical Then
comp = "Reject"
Else
comp = "Accept"
End If
If histnum = True Then
Application.Run "ATPVBAEN.XLA!Histogram", ActiveSheet.Range(ran) _
, ActiveSheet.Range(dan), ActiveSheet.Range(ban), False, False _
, False, False
End If

End Sub

' Iterative Porosity Calculations using Solver

f ttm t f tm tttm m tm m ttm ttm tttiittm ftM ttttftm m m ttff itm

A3-16
 Appendix 3: Excel Macro Listings

' 1. Simple Density assuming sandstone matrix and fluid filled pores with unknown density.
' 2. Square root porosity using density porosity and neutron porosity and adding hydrocarbon
' correction when necessary. ((PHInA2+PHIdA2)/2)Al/2
' 3. Wiley and Pratchett Porosity using density porosity from 1.
14. Gaymard and Poupon Hydrocarbon corrected Porosity using density porosity from 1.

1Water Saturation Calculations

11. Use above porosity calculations assuming that a = l, m=n=2.

Sub Iterative(Dep, Cpor, DenVal, NeutVal, WPNeutVal, ResVal, ResValXo, ResRw, _

GPDenSh, GPNeutSh, RhobSS, S a lM u d , ResSal_wat, RhobM, ResWatRhob, NumCalcPoints,
StartRow, FinRow, Namelndex)

Dim changeS, results, zeroS

Dim sqerror As Boolean
Dim b%, y%

' Insert a sheet for the calculations

Sheets.Add
ActiveSheet.Name = "Iterative Porosity"
ips = Sheets("Iterative Porosity").Name

1Set up cells for changing by Solver

Cells(5,10).Name = "RhobFluid"
Cells(2,10).Name = "DAve"
Cells(3,10).Name = "DStD"
Cells(10,10).Name = "Excavation"
Cells(8,10).Name = "NAve"
Cells(9,10).Name = "NStD"
Cells(15,10).Name = "AA"
Cells(16,10).Name = "MM"
Cells(13,10).Name = "SqAve"
Cells(14,10).Name = "SqStD"
Cells(20,10).Name = "IWPCon”
Cells(21,10).Name = "IW PNeutl"
Cells(22, 10).Name = "IWPNeut2"
Cells(23,10).Name = "IWPNeut3"
C ells(24,10).Name = "IWPDen 1"
Cells(25,10).Name = "IWPDen2"
Cells(26,10).Name = "IWPDen3"
Cells(18,10).Name = "WPAve"
Cells( 19, 10).Name = "WPStD"
Cells(31,10).Nam e = "RhobHD"
Cells(29,10).Nam e = "GPDAve"
Cells(30,10).Nam e = "GPDStD"
Cells(36, 10).Name = "RhobHN"
Cells(34,10).Nam e = "GPNAve"

A3-17
 Appendix 3: Excel Macro Listings

Cells(35,10).Nam e = "GPNStD"

' Initalise the cells for Solver

t

Range("RhobFluid") = 1
Range("AA") = 2
Range("MM") = 2
Range("Excavation") = 0.04
Range("IWPCon") = 1.916
Range("IW PNeutl”) = 0.4253
Range("IWPNeut2") = 0.001828
Range("IWPNeut3") = 0.00001885
Range("IWPDen 1") = 0.4903
Range("IWPDen2") = 0.003882
Range("IWPDen3") = 0.004397
Range("RhobHD") = 1
Range("RhobHN") = 1

' Insert the porosity formulae onto the sheet

Cells(StartRow, 1) = "-Porosity'IA " & StartRow

Cells(StartRow, 3) = "=(" & RhobSS & & ws & & Chr(64 + DenColumn) & StartRow _
& ")/(" & RhobSS & "-RhobFluid)"
Cells(StartRow, 11) = "=SQRT((" & ResRw & & ws & & Chr(64 + ResColumn) & StartRow _
& ")*(1/C" & StartRow & "A2»"

Cells(StartRow, 4) = "-Porosity'ID " & StartRow & "+ Excavation"

Cells(StartRow, 12) = "=SQRT((" & ResRw & "/"' & ws & & Chr(64 + ResColumn) & StartRow _
& ")*(1/D" & StartRow & ,,A2))"

Cells(StartRow, 5) = "=SQRT(('Porosity'!D" & StartRow & "AAA+'Porosity'!C" & StartRow & IIAMM)/2)"
Cells(StartRow, 13) = "=SQRT((" & ResRw & "/"' & ws & & Chr(64 + ResColumn) _
& StartRow & ")*(1/E" & StartRow & ,,A2))"

Cells(StartRow, 6) = "=(IWPCon" & "+(IW PNeutl" & "*'Porosity'!D" & StartRow _

& "* 100)+(IWPNeut2*('Porosityl!D" & StartRow & "*100)A2)-(IWPNeut3*(" _
& ’"Porosity'ID" & StartRow & "* 100)A3)+(IW PD enl*( 100*'Porosity'!C" _
& StartRow & "))+(IWPDen2*(100*lPorosity'!C" & StartRow _
& ")A2)-(IWPDen3*( 100*'Porosity'!C" & StartRow & ")A3/100))/100"
Cells(StartRow, 14) = "=SQRT((" & ResRw & & ws & & Chr(64 + ResColumn) & StartRow _
& ")*(1/F" & StartRow & ”A2))"

Cells(StartRow, 7) = "=(" & GPConl & "*'Porosity'!C" & StartRow & ")*((" & RhobSS & ”-(l-" _
& "SQRT(" & RMF & & ws & & Chr(64 + ResXoColumn) & StartRow _
& "*'Porosity'!C" & StartRow & "A 2)))*(" & GPCon2 + GPCon3 * Sal_Mud & ")-" _
& "SQRT(" & RMF & & ws & "'!" & Chr(64 + ResXoColumn) & StartRow _
& "*,Porosity'!C" & StartRow & "A 2))*RhobHD*(1.33-0.036*RhobHD+0.2*RhobHDA2)-(" _
& GPCon3 & "*K" & StartRow & ")*(" & ResSal_wat - Sal_Mud & "))/(" _
& RhobSS & "-(" & RhobM & "*(1-K" & StartRow _
& ”-SQRT(" & RMF & & ws & '"!" & Chr(64 + ResXoColumn) & StartRow _

A3-18
 Appendix 3: Excel Macro Listings

& "*'Porosity'!C" & StartRow & "A 2))" & ")+(RhobHD*" _

& "SQRT(" & RMF & & ws & & Chr(64 + ResXoColumn) & StartRow _
& "*'Porosity'!C" & StartRow & "A 2))+" & ResWatRhob & "*K" & StartRow & "))))"
Cells(StartRow, 15) = "=SQRT((" & ResRw & & ws & & Chr(64 + ResColumn) & StartRow
& ")*(1/G" & StartRow & "A2))n

Cells(StartRow, 8) = "-Porosity'ID " & StartRow & "-(-('Porosity'ID" & StartRow _

& & "SQRT(" & RMF & & ws & '"!" & Chr(64 + ResXoColumn) _
& StartRow & '""'Porosity'ID" & StartRow & "A 2))" _
& "*RhobHN)*(((1.97-0.324*RhobHN*+1.8*RhobHNA2-" & GPCon4 * S a lM u d _
& ")/(l-" & GPCon4 * Sal Mud & "))+" _
& '"Porosity'ID" & StartRow & "*L" & StartRow & "*((" & GPCon4 _
& "*(" & Sal_Mud - ResSal wat & "))/(l-" & GPCon4 * Sal_Mud & ")))"
Cells(StartRow, 16) = "=SQRT((" & ResRw & "/"' & ws & '"!" & Chr(64 + ResColumn) & StartRow
& ")*(1/H" & StartRow & ”A2))"
Range(Cells(StartRow, 1), Cells(FinRow, 8)).Select
Selection.FillDown
Range(Cells(StartRow, 11), Cells(FinRow, 16)).Select
Selection.FillDown

For j = StartRow To FinRow

If Not IsEmpty(Cpor(j - 1)) Then
Cells(j, 2) = "-Porosity'IB " & j
Cells(j, 18) = "=(C" & j & "-B" & j & ")"
Cells(j, 19) = "=(D" & j & "-B" & j & ")"
Cells(j, 20) = "=(E" & j & "-B" & j & ")"
Cells(j, 21) = "=(F" & j & "-B" & j & ")"
Cells(j, 22) = "=(G" & j & "-B" & j & ")"
Cells(j, 23) = "=(H" & j & "-B" & j & ")"
End If
N extj

' Find average, standard deviation difference between calculated and core porosity
' Also write names to cells

Namelndex(16) = "IDepth"
Namelndex(17) = "ICPor"
Namelndex(18) = "IDPor"
Namelndex(19) = "INPor"
Namelndex(20) = "ISPor"
Namelndex(21) = "IWPPor"
Namelndex(22) = "IGPDPor"
Namelndex(23) = "IGPNPor"

Namelndex(26) = "ISwDPor"
Namelndex(27) = "ISwNPor"
Namelndex(28) = "ISwSPor”
Namelndex(29) = "ISwWPPor"
Namelndex(30) = "ISwGPPor"
Namelndex(31) = "ISwGPPor"

A3-19
 Appendix 3: Excel Macro Listings

Cells(l, 1) = "Depth"
Cells(l, 2) = "Core Porosity"
Cells(l, 3) = "Density Porosity"
Cells(l, 4) = "Neutron Porosity"
Cells(l, 5) = "Square Root Porosity"
Cells(l, 6) = "Wiley & Pratchett Porosity"
Cells(l, 7) = "Gaymard & Poupon Density Porosity"
Cells(l, 8) = "Gaymard & Poupon Neutron Porosity"

Cells(l, 11) = "Sw Density Porosity"

Cells(l, 12) = "Sw Neutron Porosity"
Cells(l, 13) = "Sw Square Root Porosity"
Cells(l, 14) = "Sw Wiley & Pratchett Porosity"
Cells(l, 15) = "Sw Gaymard & Poupon Density Porosity"
Cells(l, 16) = "Sw Gaymard & Poupon Neutron Porosity"

Cells(l, 18) = "Density - Core Porosity"

Cells(l, 19) = "Neutron - Core Porosity"
Cells(l, 20) = "Square Root - Core Porosity"
Cells(l, 21) = "Wiley & Pratchett - Core Porosity"
Cells(l, 22) = "Gaymard & Poupon Density - Core Porosity"
Cells(l, 23) = "Gaymard & Poupon Neutron - Core Porosity"

Cells(2,9) = "Density - Core Porosity Average"

Cells(3,9) = "Density - Core Porosity Standard Deviation"
Cells(5,9) = "Rhob Fluid"
Cells(2,10) = "=AVERAGE(R" & StartRow & ":R" & FinRow & ")"
Cells(3,10) = "=STDEV(R" & StartRow & ":R" & FinRow & ")"

Cells(8,9) = "Neutron - Core Porosity Average"

Cells(9,9) = "Neutron - Core Porosity Standard Deviation"
Cells(10,9) = "Excavation"
Cells(8, 10) = "=AVERAGE(S" & StartRow & ":S" & FinRow & ")"
Cells(9, 10) = "=STDEV(S" & StartRow & ":S" & FinRow & ")"

Cells(13,9) = "Square Root - Core Porosity Average"

Cells(14, 9) = "Square Root - Core Porosity Standard Deviation"
Cells(15,9) = "Neutron Exponent"
Cells(16,9) = "Density Exponent"
Cells(13,10) = "=AVERAGE(T" & StartRow & ":T" & FinRow & ")"
Cells(14,10) = "=STDEV(T" & StartRow & ":T" & FinRow & ")"

Cells(18,9) = "Wiley & Pratchett - Core Porosity Average"

Cells(19,9) = "Wiley & Pratchett - Core Porosity Standard Deviation"
Cells(20,9) = "Constant"
Cells(21,9) = "Neutron 1 Constant"
Cells(22,9) = "Neutron2 Constant"
Cells(23,9) = "Neutron3 Constant"
Cells(24, 9) = "Density 1 Constant"

A3-20
 Appendix 3: Excel Macro Listings

Cells(25,9) = "Density2 Constant"

Cells(26,9) = "Density3 Constant"
Cells(18,10) = "=AVERAGE(U" & StartRow & ":U" & FinRow & ")"
Cells(19,10) = "=STDEV(U" & StartRow & ":U" & FinRow & ")"
I

Cells(29,9) = "Gaymard & Poupon Density - Core Porosity Average"

Cells(30, 9) = "Gaymard & Poupon Density - Core Porosity Standard Deviation"
Cells(31,9) = "Hydrocarbon Density"
Cells(29,10) = "=AVERAGE(V" & StartRow & ":V" & FinRow & ")"
Cells(30,10) = "=STDEV(V" & StartRow & ":V" & FinRow & ")"

Cells(34,9) = "Gaymard & Poupon Neutron - Core Porosity Average"

Cells(35,9) = "Gaymard & Poupon Neutron - Core Porosity Standard Deviation"
Cells(36,9) = "Hydrocarbon Density"
Cells(34,10) = "=AVERAGE(W" & StartRow & ":W" & FinRow & ")"
Cells(35,10) = "=STDEV(W" & StartRow & ":W" & FinRow & ")"

' Pass cell references for iteration to Solver to minimise standard deviation on
' condition that the difference with is = 0.

For k = 1 To 6
If k = 1 Then
zero = "DAve"
change = "RhobFluid"
result = "DStD"
Call PorSolver(result, change, zero)
Elself k = 2 Then
zero = "NAve"
change = "Excavation"
result = "NStD"
Call PorSolveifresult, change, zero)
Elself k = 3 Then
zero = "SqAve"
change = "AA, MM"
result = "SqStD"
Call PorSolver(result, change, zero)
For j = 1 To NumCalcPoints
b=j + 1
sqerror = Application. WorksheetFunction.IsError(Cells(b, 4))
If sqerror = True Then
Cells(b, 4) = 0
Cells(b, 10) = 0
End If
Nextj
Elself k = 4 Then
zero = "WPAve"
change = "IWPNeutl, IWPNeut2, IWPNeut3, IW PDenl, IWPDen2, IWPDen3"
result = "WPStD"
Call PorSolver(result, change, zero)
Elself k = 5 Then

A3-21
 Appendix 3: Excel Macro Listings

zero = "GPDAve"
change = "RhobHD"
result *= "GPDStD"
Call PorSolver(result, change, zero)
Elself k = 6 Then
zero = "GPNAve"
change = "RhobHN"
result = "GPNStD"
Call PorSolver(result, change, zero)
End If
Nextk

Range(Cells(2,1), Cells(FinRow, 18)).NumberFormat = "0.000"

For y = 2 To NumCalcPoints + 1
If IsNumeric(Cells(y, 5)) = False Then
Cells(y, 5) = 0
Cells(y, 13) = 0
End If
Nexty

End Sub

m f m m f m tm titm tf i in m tm ttf im tM m tm if m m t ttm m m ttf

1 Solver Routine '

Sub PorSolver(result, change, zero)

SolverReset
SolverOptions precision:=0.0001
SolverOk SetCell:=Range(result), MaxMinVal:=2, ByChange:=Range(change)
SolverAdd cellRef:=Range(zero), relation:=2, formulaText:=0
SolverSolve userFinish:=True
SolverFinish keepFinal:=l
End Sub

tlttfttttlltttlltV ttltllttttllfttlltlttV tlltfttltlttlM ltlltlllM IV V IIttfttlt

' Plot Routine '

ftititiittttitiim tftiitttm ttH ffttttM tfm tftiM ttttfM tfm fitttiftfttt

1Most of this routine is adapted from J Walker's xyscatter routine from ’

1his web site: http://www.j-walk.com/ss/excel/index.htm 1
I t

m m tttf H im ttt im m t m tm m t itti ittm tm ttm if tiim m tim t iii

Sub PlotGraph(ws, ps, ips, ItPlot, StartRow, FinRow, MaxCpor, Namelndex)

Dim NumCols%, XVarIndex%, rl$

Dim q%, t%, v2!, v4!, ct%
Dim v l, v3, PlotVarlndex(lOO) As Integer

A3-22
 Appendix 3: Excel Macro Listings

Dim drr, dm, r2 As Range

Dim XTit$, XTitNS, YTit$, YTitN$, Title$, TitleN$
Dim R%
t

SourceDataSheet = ps
Sheets(ps).Select
7 For k = 1 To 2
If k = 1 Then
m = 1

s=7
NumCols = 8
If SourceDataSheet = ips Then
YTit = "Iteratively Calculated Porosity"
YTitN = "Iter Calc Por"
Else
YTit = "Calculated Porosity"
YTitN = "Calc Por"
End If
Elself k = 2 Then
NumCols = 7
If SourceDataSheet = ips Then
YTit = "Iteratively Calcalted Sw"
YTitN = "Iter Calc Sw"
m = 10
s = 15
Else
YTit = "Calculated Sw"
YTitN = "Calc Sw"
m = 10
s = 14
End If
End If
t = s+ 1
For i = m To s
If i = 2 And CporColumn = 0 Then GoTo 11
Sheets(SourceDataSheet).Select
XTit = Sheets(SourceDataSheet).Cells(l, i).Text
q= i+ 1
XVarlndex = i
If SourceDataSheet = ips And i = 10 Then
drr = Range(Cells(StartRow, 1), Cells(FinRow, 1))
v3 = Application.WorksheetFunction.Min(drr)
v4 = Application. WorksheetFunction.Max(drr)
ct = xlXYScatterLinesNoMarkers
E lself i = 1 Or i = 10 Then
Set drr = Range(Cells(StartRow, 1), Cells(FinRow, 1))
v3 = Application.WorksheetFunction.Min(drr)
v4 = Application. WorksheetFunction.Max(drr)
ct = xlXYScatterLinesNoMarkers
Else

A3-23
 Appendix 3: Excel Macro Listings

v3 = 0
ct = xlXYScatter
If i = 2 Then
v4 = MaxCpor
Else
Set drr = Range(Cells(StartRow, i), Cells(FinRow, i))
v4 = Application.WorksheetFunction.Max(drr)
End If
v4 = 0.05 * (Int(v4 / 0.05) + 1)
If v4 >= 1 Then
v4 = 1
End If
End If
If SourceDataSheet = ips And i = 10 Then
XVarlndex = 1
Elself i = 10 Then
XVarlndex = 1
Else
XVarlndex = i
End If
9 For j = i + 1 To s + 1
PlotVarIndex(j) = j
Next
vl = 0
Set r2 = Range(Cells(StartRow, q), Cells(FinRow, t))
If q = 2 Then
v2 = Application.WorksheetFunction.Max(Range(Cells(StartRow, q + 1), Cells(FinRow, t)))
If v2 < MaxCpor Then
v2 = MaxCpor
End If
Else
v2 = Application.WorksheetFunction.Max(r2)
End If
v2 = 0.05 * (Int(v2 / 0.05) + 1)
If v2 >= 1 Then
v2= 1
End If
If v2 <= 1 And v4 <= 1 Then
If v2 >= v4 Then
v4 = v2
Else
v2 = v4
End If
End If
I f k= 1 Then
If i = 1 And CporColumn = 0 Then
rl = Chr(64 + XVarlndex) & "1:" & Chr(64 + XVarlndex) & FinRow & & Chr(64 + m + 2) & "1:" & Chr(64 + s + 1)
& FinRow
Else
rl = Chr(64 + XVarlndex) & "1:" & Chr(64 + PlotVarIndex(NumCols)) & FinRow

A3-24
 Appendix 3: Excel Macro Listings

End If
Elself k = 2 Then
If i = 10 Then
If XVarlndex = 1 Then
rl = Chr(64 + XVarlndex) & "1:" & Chr(64 + XVarlndex) & FinRow & & Chr(64 + m) & "1:" & Chr(64 + s + 1) &
FinRow
Else
rl = Chr(64 + XVarlndex) & "1 & Chr(64 + PlotVarIndex(s + 1)) & FinRow
End If
Elself i > 10 Then
rl = Chr(64 + XVarlndex) & "1 & Chr(64 + PlotVarIndex(s + 1)) & FinRow
End If
If SourceDataSheet = ips And XVarlndex = 1 Then
rl = Chr(64 + XVarlndex) & "1:" & Chr(64 + XVarlndex) & FinRow & & Chr(64 + m + 1) & "1:" & Chr(64 + s + 1)
& FinRow
End If
End If
If XVarlndex = 1 Then
XTit = Sheets(SourceDataSheet).Cells(l, l).Text
XTitN = Namelndex(l)
Elself XVarlndex =10 Then
XTit = Sheets(SourceDataSheet).Cells(l, 10).Text
XTitN = Namelndex(lO)
Elself SourceDataSheet = ips And XVarlndex =10 Then
XTit= Sheets(SourceDataSheet).Cells(l, 10).Text
XTitN = NameIndex( 10)
Else
XTit = Sheets(SourceDataSheet).Cells(l, i).Text
If SourceDataSheet = ips Then
XTitN = Namelndex(15 + i)
Else
XTitN = Namelndex(i)
End If
End If
Title = XTit & " v . " & YTit
TitleN = XTitN & " v. " & YTitN
Call MakePlot(XVarIndex, SourceDataSheet, rl, v l, v2, v3, v4, ct, Title, TitleN, XTit, YTit)
If i = 10 Then
If SourceDataSheet = ips Then GoTo 11
If XVarlndex = 10 Then GoTo 11
XVarlndex = 10
q = i+ 1
Sheets(SourceDataSheet). Select
v3 = 0
ct = xlXYScatter
Set drr = Range(Cells(StartRow, i), Cells(FinRow, i))
v4 = Application.WorksheetFunction.Max(drr)
v4 = 0.05 * (Int(v4 / 0.05) + 1)
If v4 >= 1 Then
v4 = 1

A3-25
 Appendix 3: Excel Macro Listings

End If
GoTo 9
End If
11 Next
Next

If ItPlot = True Then

SourceDataSheet = ips
ItPlot = False
ItPlot2 = True
GoTo 7
End If

' Cumulative Frequency Plots

If Stat = True Then

200 XTit = "Porosity"
YTit = "Frequency"
Title = "Cumulative Frequency Plot"
If SourceDataSheet = ps Then
Sheets(ps).Select
Title = Title & " using assumed constants"
TitleN = "Cum Freq Porosity Plot"
If CporColumn = 0 Then
rl = "AB19:AH" & (20 + histstop)
Else
rl = "AA19:AH" & (20 + histstop)
End If
Else
Sheets(ips).Select
Title = Title & " using constants calculated iteratively"
TitleN = "Cum Freq Iter Por Plot"
rl = "AI19:AP" & (20 + histstop)
End If
ct = xlXYScatterLines
vl = 0
v2 = 1
v3 = 0
v4 = MaxCpor
v4 = 0.05 * (Int(v4 / 0.05) + 1)
If v4 >= 1 Then
v4 = 1
End If
Call MakePlot(XVarIndex, SourceDataSheet, rl, v l, v2, v3, v4, ct, Title, TitleN, XTit, YTit)
If ItPlot2 = True Then
SourceDataSheet = ps
ItPlot2 = False
GoTo 200
End If
End If

A3-26
 Appendix 3: Excel Macro Listings

End Sub

iti m iii im f f r n tf tm m iiitf tm ttim t m m f tm m i m m tf f im ii m tf tm tm f i

' Plot variables!

Sub MakePlot(XVarIndex, SourceDataSheet, rl, v l, v2, v3, v4, ct, Title, TitleN, XTit, YTit)

Dim U%, cou%, NewPlot%

Dim OneLine As Range
I

If Not ct = xlXYScatterLines Then

NewPlot = 2 * XVarlndex
Sheets(SourceDataSheet).Select
Cells(10,40 + NewPlot).Value = vl
Cells( 11,40 + NewPlot).Value = v2
Cells(10,41 + NewPlot).Value = v3
Cells(l 1,41+ NewPlot) Value = v4
Set OneLine = Sheets(SourceDataSheet).Range(Cells(10,40 + NewPlot), Cells(l 1, 41+ NewPlot))
End If

Charts.Add
With ActiveChart
.ChartType = ct
.SetSourceData Source:=Sheets(SourceDataSheet).Range(rl), PlotBy:=xlColumns
.Location xlLocationAsNewSheet
.HasLegend = True
.HasTitle = True
.ChartTitle.Text = wks & & Title
.Name = TitleN
With .PlotArea
.Border.Weight = xlThin
.Border.LineStyle = xlAutomatic
.Interior.Colorlndex = xlNone
.Width = Application.UsableWidth
.HEIGHT = Application.UsableHeight
End With
With .Axes(xlValue)
.MinimumScale = vl
.MaximumScale = v2
.MinorUnitlsAuto = True
.MajorUnitlsAuto = True
.Crosses = xlAutomatic
.ReversePlotOrder = False
.ScaleType = xlLinear
.HasTitle = True
.AxisTitle.Text = YTit
.HasMajorGridlines = True
End With

A3-27
 Appendix 3: Excel Macro Listings

With .Axes(xlCategory)
.MinimumScale = v3
.MaximumScale = v4
.MinorUnitlsAuto = True
.MajorUnitlsAuto = True
.Crosses = xlAutomatic
.ReversePlotOrder = False
.ScaleType = xlLinear
.HasTitle = True
.AxisTitle.Text = XTit
.HasMajorGridlines = True
End With
With .Legend
.Left =100
.Top = 50
.Width = 170
.Font.Size = 8
With .Border
.Weight = xlHairline
.LineStyle = xlNone
End With
.Shadow = False
.Interior.Colorlndex = xlNone
End With
If .ChartType = xlXYScatter Then
With .PageSetup
.TopMargin = .LeftMargin
.BottomMargin = .LeftMargin
.RightMargin = Application.InchesToPoints(l 1.69 - _
(.LeftMargin / 72 + (8.27 - 2 * (.LeftMargin / 72))))
End With
With .PlotArea
.Top = 20
.HEIGHT = 440
.Width = .HEIGHT
End With
cou = .SeriesCollection.count
For U = .SeriesCollection.count To 1 Step -1
With .SeriesCollection(U)
.MarkerStyle = xlX
.MarkerBackgroundColorlndex = xlNone
.MarkerSize = 5
If cou - U = 0 Then
.MarkerForegroundColorlndex =10
Elself cou - U = 2 Then
.MarkerForegroundColorlndex = 13
Else
.MarkerForegroundColorlndex = cou - U
End If
End With

A3-28
 Appendix 3: Excel Macro Listings

Next
OneLine.Copy
.SeriesCollection.Paste Rowcol:=xlColumns, SeriesLabels:=False,
CategoryLabels:=True, Replace:=False, NewSeries:=True
Application.CutCopyMode = False
With .SeriesCollection(cou + 1)
.MarkerStyle = xlNone
.Name = "=""l to 1 line"""
With .Border
.Colorlndex = 1
.Weight = xlHairline
.LineStyle = xlContinuous
End With
End With
.ChartArea. Select
ActiveWindow.Zoom = 87
Elself .ChartType = xlXYScatterLines Then
With .PageSetup
.TopMargin = .LeftMargin
.BottomMargin = .LeftMargin
.RightMargin = Application.InchesToPoints(l 1.69 - _
(.LeftMargin / 72 + (8.27 - 2 * (.LeftMargin / 72))))
End With
With .PlotArea
.Top = 30
.HEIGHT = 440
.Width = .HEIGHT
End With
cou = .SeriesCollection.count
For U = .SeriesCollection.count To 1 Step -1
With .SeriesCollection(U)
.MarkerStyle - xlX
.MarkerBackgroundColorlndex = xlNone
.MarkerSize = 5
If cou - U = 0 Then
.MarkerForegroundColorlndex = 10
.Border.Colorlndex = 10
Elself cou - U = 2 Then
.MarkerForegroundColorlndex = 13
.Border.Colorlndex =13
Else
.MarkerForegroundColorlndex = cou - U
.Border.Colorlndex = cou - U
End If
End With
Next
With .Legend
.Left = 290
.Top = 333
.Font.Size = 8

A3-29
 Appendix 3: Excel Macro Listings

End With
.ChartTitle.Font.Size =10
.Axes(xlValue).MajorUnit = 0.2
.ChartArea.Select
ActiveWindow.Zoom = 87
Elself .ChartType = xlXYScatterLinesNoMarkers Then
.ChartArea.Select
ActiveWindow.Zoom = 94
cou = .SeriesCollection.count
For U = .SeriesCollection.count To 1 Step -1
With .SeriesCollection(U)
If cou - U = 6 Then
.MarkerStyle = xlX
.MarkerBackgroundColorlndex = xlNone
.MarkerForegroundColorlndex = 9
.MarkerSize = 5
.Border.LineStyle = xlNone
Elself cou - U = 2 Then
.Border.Colorlndex = 13
Elself cou - U = 0 Then
.Border.Colorlndex =10
Else
.Border.Colorlndex = cou - U
End If
End With
Next
With .ChartTitle
.Orientation = xlUpward
.Left = xlLeft
.Top = 100
End With
With .PlotArea
.Top = 0
.HEIGHT = 430
.Width = 670
.Left =55
End With
With .Axes(xlValue).TickLabels
.NumberFormat = "0.0"
.Orientation = xlUpward
End With
With .Axes(xlCategory).TickLabels
.NumberFormat = "0"
.Orientation = xlUpward
End With
With .Axes(xlCategory)
.AxisTitle.Orientation = xlUpward
.AxisTitle.Top = 670
End With
With .Legend

A3-30
 Appendix 3: Excel Macro Listings

.Left = 114
.Top = 10
.Font.Size = 8
End With
End If
End With

End Sub
t t t m m tf tim tttm m m m m t ttii pnH

A3-31
 Bibliography

Bibliography

Alberty, M. (1996). “The influence o f the borehole environment upon

compressional sonic logs.” The Log AnalvstfJuly-Augusf).
Allen, D., et al. (1997). “How to use Borehole Nuclear Magnetic Resonance.”
Oilfield ReviewfSummer): 34-57.
Allen, D. F., et al. (1990). “The Effect o f Wellbore condition on Wireline and
MWD Neutron Density Logs.” 65th Annual Technical Conference and Exhibition o f
the Society o f Petroleum Engineers SPE 20563.
Allen, L. S., et al. (1967). “Dual-spaced neutron logging for porosity.”
Geophysics 32: 60-68.
Analysts, SPWLA (1997). Glossary o f Well Logging Terms. Houston, Society
o f Professional Well Log Analysts.
Anderson, B., et al. (1996). The response o f multiarray induction tools in
highly dipping formations with invasion and in arbitrary 3D geometries. 37th Annual
Logging Symposium Transactions, Society o f Professional Well Log Analysts.
Anderson, B., et al. (1990). Response o f 2-MHz LWD resistivity and wireline
induction tools in dipping beds and laminated formations. 31st Annual Logging
Symposium Transactions, Society o f Professional Well Log Analysts.
Anderson, B., et al. (1992). “Response o f 2-MHz LWD resistivity and wireline
induction tools in dipping beds and laminated formations.” The Log Analyst 33(5):
461-475.
Anderton, R. (1993). Sedimentation and basin evolution in the Paleogene o f
the Northern North Sea and Faeroe-Shetland basins. Petroleum Geology o f Northwest
Europe: Proceedings o f the 4th Conference. J. R. Parker. London, The Geological
Society: 31.
Archie, G. E. (1942). “The electrical resistivity log as an aid in determining
some reservoir characterists.” Petroleum Technology^January): 4-13.
Archie, G. G. (1942). “Electrical resistivity as an aid in core analysis
interpretation.” Trans AIME 146: 54-63.

B -l
 Bibliography

Armentrout, J. M., et al. (1993). Log-motif analysis o f Palaeogene

depositional systems tracts, Central and Northern North Sea: defined by sequence
stratigraphic analysis. Petroleum Geology o f Northwest Europe: Proceedings o f the
4th Conference. J. R. Parker. London, The Geological Society: 45-57.
Arnold, D. M., et al. (1981). Experimental determination o f environmental
corrections for a Dual-Spaced Neutron Porosity Log. 22nd Annual Logging
Symposium Transactions, Society o f Professional Well Log Analysts.
Austin, S. P., et al. (1994). The observation o f anomalous sonic log responses
in deviated wells in the southern North Sea. 16th European Formation Evaluation
Symposium Transactions, Aberdeen Chapter, Society o f Professional Well Log
Analysts.
Banner, J. A., et al. (1992). Guillemot D: a successful appraisal through
alternative interpretation. Exploration Britain. R. F. P. Hardman. London, The
Geological Society. The Geological Society Special Publication No. 67: 129-149.
Barber, D. J. L. (1989). An introduction to the properties o f condensed matter.
Cambridge, Cambridge University Press.
Bateman, R. M. (1985). Qpen-hole log analysis and formation evaluation.
Boston, Reidel Publishing Company.
Bedford, J., et al. (1997). The empirical investigation o f density anisotropy in
horizontal gas w ells. 38th Annual Logging Symposium Transactions, Society o f
Professional Well Log Analysts.
Best, D., et al. (1990). An innovative approach to correct density
measurements while drilling for hole size effect. 31st Annual Logging Symposium
Transactions, Society o f Professional Well Log Analysts.
Betts, P., et al. (1990). “Acquiring and Interpretating logs in Horizontal
W ells.” Oilfield Review 2f3L 34-51.
Bigelow, E. L., et al. (1992). “A New Frontier: Log Interpretation in
Horizontal Wells.” SPWLA 33rd Annual Logging Symposium Transactions 1: O O l-
006.
Bonner, S., et al. (1996). “Resistivity While Drilling-Images from the String.”
Oil Review Spring 1996.
Bourbie, T., et al. (1987). Acoustics o f porous media. Houston, Gulf
Publishing Company.

B-2
 Bibliography

Bourgeois, T. J., et al. (1998). “Pushing the Limits o f Formation Evaluation

While Drilling.” Oilfield Review W inter 1998.
Bowman, M. B. J. (1998). Cenozoic. Introduction to the Petroleum Geology o f
the North Sea: Basic concepts and recent advances. K. W. Glennie. Oxford, Blackwell
Science Ltd: 350-375.
Brami, J. B. (1991). “Current Calibration and Quality Control practices for
Selected Measurement while drilling Tools.” 66th Annual Technical Conference and
Exhibition o f the Society o f Petroleum Engineers SPE 22540.
Brooks, J., et al. (1987). Petroleum Geology o f North West Europe:
Proceedings o f the 3rd Conference on Petroleum Geology o f North West Europe.
Petroleum Geology o f North West Europe: Proceedings o f the 3rd Conference on
Petroleum Geology o f North West Europe. J. Brooks and K. W. Glennie. London,
Graham & Trotman. 1.
Bulat, J., et al. (1987). Uplift determination from interval velocity studies, UK
southern North Sea. Petroleum Geology o f North West Europe: Proceedings o f the
3rd Conference on Petroleum Geology o f North West Europe. J. Brooks and K. W.
Glennie. London, Graham & Trotman: 293-305.
Burnett, T. M., Koopersmith, C. A. and Spross, R.L. (1990). “Drill Collar
effects on MWD epithermal, thermal and capture gamma ray neutron porosity
measurements.” SPWLA 31st Annual Logging Symposium Transcations 1: E1-E22.
Cannon, D. (1998). “Quantification o f Hydrocarbon saturation o f Borehole
imaging: All may not be as it seems.” SPWLA 39th Annual Logging Symposium
Workshop.
Cannon, D. E., et al. (1999). Interpretation o f asymmetrically invaded
formations with azimuthal and radial LWD data. 40th Annual Logging Symposium
Transactions, Society o f Professional Well Log Analysts.
Carpenter, W., et al. (1997). Applications and interpretation o f azimuthallv
sensitive density measurements acquired while drilling. 38th Annual Logging
Symposium Transactions, Society o f Professional Well Log Analysts.
Chen, S. T. (1988). “Shear-wave logging with dipole sources.” Geophysics 53:
659-667.
Cloetingh, S., et al. (1987). Apparent sea-level fluctuations and a palaeostress
field for the North Sea region. Petroleum Geology o f North West Europe:

B-3
 Bibliography

Proceedings o f the 3rd Conference on Petroleum Geology o f North West Europe. J.

Brooks and K. W. Glennie. London, Graham & Trotman. 1: 49-57.
Coope, D. F. (1983). “Gamma ray measurement while drilling.” The Log
Analyst 24(1): 3-9.
Cowan, P., et al. (1997). Investsigations into anomalous responses o f gamma
density and neutron porosity tools in horizontal gas w ells. 38th Annual Logging
Symposium Transactions, Society o f Professional Well Log Analysts.
Cuddy, S. (1997). The application o f the mathematics o f fuzzzv logic to
petrophvsics. 38th Annual Logging Symposium Transactions, Society o f Professional
Well Log Analysts.
Cuddy, S., et al. (1994). The observed effects o f high deviation wells on
porosity measurements—a discussion o f potential explanations. 16th European
Formation Evaluation Symposium Transactions, Aberdeen Chapter, Society o f
Professional Well Log Analysts.
Cunningham, A. B., Bullion, D. and Opstad, E.A. (1990). “Use o f MWD
Formation Evaluation in the Endicott Reservoir North Slope, Alaska, USA.”
Thirteenth European Formation Evaluation Symposium Transactions 1: C1-C21.
Cunningham, A. B., and Opstad, E.A. (1992). “Use o f MWD Formation
Evaluation in the Endicott Reservoir North Slope, Alaska, USA.” The Log
Analvstf September-October!: 439-450.
Czubek, J. (1960). “Physical possiblities o f Gamma-Gamma Logging.”
Geophysics 37: 249-275.
Day, S. N. J., et al. (1990). Tool motion and borehole environmental effects on
MWD neutron porosity. 13th European Formation Evaluation Symposium
Transactions, Budapest Chapter, Society o f Professional Well Log Analysts.
Delafor, H. (1984). “Prediction software improves directional drilling.” World
Qil 189(April).
Den Hartog Jager, D., et al. (1993). Evolution o f Palaeogene submarine fans
o f the North Sea in space and time. Petroleum Geology o f Northwest Europe:
Proceedings o f the 4th Conference. J. R. Parker. London, The Geological Society: 59-
71.
Desbrandes, D. (1985). Encyclopedia o f Well Logging. Paris, Editions
Technip.

B-4
 Bibliography

Dorn, G., et al. (1996). “Geophysical reservoir characterization o f Pickerill

Field, North Sea, using 3-D seismic and well data.” AAPG Studies in Geology 42:
107-122.
Doveton, J. H. (1994). G eologic Log Analysis using computer methods. Tulsa,
The American Association o f Petroleum Geologists.
Dullien, F. A. L. (1992). Porous Media: Fluid transport and pore structure.
London, Academic Press , Inc.
Dvorkin, J., et al. (1998). “Time-average equation revisited.” Geophysics
63(2): 460-464.
Ellis, D., et al. (2000). “Interpretating neutron logs in horizontal wells - a
forward modelling tutorial.” Petrophvsics: 23-32.
Ellis, D. V. (1986). “Neutron porosity devices-what do they measure?” First
Break 4(1): 11-17.
Ellis, D. V. (1987). Well Logging for Earth Scientists. N ew York, Elsevier.
Elphick, R. Y., et al. (1996). “A deterministic method for calculating facies
from wireline logs.” The Log Analyst 37(March-April): 14-25.
Evans, M., et al. (1999). A n ovel approach for compensating neutron porosity
logs for borehole effects. 40th Annual Logging Symposium Transactions, Society o f
Professional Well Log Analysts.
Evans, M., et al. (1988). “Formaton porosity measurement while drilling.”
SPWLA 29th Annual Logging Sym posium Transactions 1: C1-C18.
Fritz, R. D. (1991). HD Definitions, History and Classifications. Geological
Aspects o f Horizontal Drilling. R. D . Fritz, Horn, M.K. and Joshi, S.D. Tulsa,
Oklahoma, AAPG. 33: 1-51.
Galloway, W. E., et al. (1 993). Sequence stratigraphic and depostional
framework o f the Cenozoic fill, Central and Northern North Sea basins. Petroleum
Geology o f Northwest Europe: Proceedings o f the 4th Conference. J. R. Parker.
London, The Geological Society: 33-43.
Gassmann, F. (1951). “Uber die elastizitat poroser medien.” Vierteljahrsschrift
der Naturforschenden Gesellschaft in Zurich 96: 1-23.
Gaymard, R., et al. (1968). “R esponse o f neutron and formation density logs
in hydrocarbon bearing formations.” T he Log Analyst 9(5): 3-12.

B-5
 Bibliography

Gianzero, S., et al. (1992). Induction, resistivity and MWD tools in horizontal
wells. Resistivity Reprint Volume. Society o f Professional Well Log Analysts: paper
30, 1-14.
Glennie, K. (1990). Outline o f North Sea History and Structural Framework.
Introduction to the Petroleum Geology o f the North Sea. K. Glennie. London,
Blackwell. 1: 34-76.
Glennie, K. W. (1990). Upper Permian - Rotliegend. Introduction to the
Petroleum Geology o f the North Sea. K. W. Glennie. London, Blackwell Scientific
Publications: 120-152.
Goodway, B. (2001). “AVO and Lame constants for rock parameterization
and fluid detection.” Canadian Society o f Explorational Geophysics Recorder 26(6):
39-60.
Han, D.-H., et al. (1986). “Effects o f porosity and clay content on the wave
velocities in sandstones.” Geophysics 51: 2093-2107.
Hansen, P., et al. (1996). Unraveling the difference between LWD and
wireline measurements. 37th Annual Logging Symposium Transactions, Society o f
Professional Well Log Analysts.
Harris, R., et al. (1993). “Time Lapse Analysis o f Logging While Drilling and
Wireline Data: Invasion Radius and Rt from Forward and Inverse Modelling.”
SPWLA 34th Annual Logging Symposium Transactions 2: MM1-MM13.
Harvey, B. G. (1969). Introduction to Nuclear Phvsics and Chemistry. New
Jersey, Prentice-Hall Inc.
Harvey, P. K. (1999). Ocean Drilling Program coring policy.
Harvey, P. K., et al. (1992). “Mineral transforms and downhole geochemical
measurements.” Scientific Drilling 1(4): 163-176.
Hilchie, D. W. (1982). Applied Openhole Log Interpretation for Geologists
and Engineers. Golden, Colorado, Douglas W. H ilchie Inc.
Hodgson, N. A., et al. (1992). Salt-realted tectonics, sedimentation snd
hydrocarbon plays in the Central Graben, North Sea, UKCS. Exploration Britain. R.
F. P. Hardman. London, The Geological Society. T he G eological Society Special
Publication No. 67: 31-63.
Holbrook, P., et al. (1999). Velocity-Porositv-Mineralogy Gassmann
coefficient mixing relationships for water saturated sedimentary rocks. 40th Annual
Logging Symposium Transactions, Society o f Professional W ell Log Analysts.

B-6
 Bibliography

Holenka, J., et al. (1995). Azimuthal Porosity while Drilling. 36th Annual
Logging Syposium Transactions, Society o f Professional Well Log Analysts.
Howarth, S. (1997). A Century in Oil: The "Shell" Transport and Trading
Company 1897-1997. London, Weidenfeld & Nicolson.
Hurst, A. (1990). Natural gamma-ray spectrometry in hydrocarbon-bearing
sandstones from the Norweigian Contential Shelf. Geological Applications o f
Wireline Logs. A. Hurst, M. A. Lovell and A. C. Morton. London, The Geological
Society. Geological Society Special Publications No. 48: 211-222.
Hutchinson, M., et al. (1991). Compatibility o f Formation Evaluation MWD
data and wireline log data (Guidelines to optimise FEMWD quality: IMS. Eastern
Hemisphere Technology Committee). Fourteenth European Formation Evaluation
Symposium.
Hutchinson, M., et al. (1996). Optimizing MWD Operations (Recommended
Practices and Operational Guideline). IMS.
INTEQ, B. H. (1999). MWD Brochure http://www.bakerhughes.com/
inteq/Evaluation/mwdshelfset-htm. Houston, Baker Hughes INTEQ.
Jackson, C. E., et al. (1994). “Improving Formation Evaluation by Resolving
Differences between LWD and Wireline Log data.” 69th Annual Technical
Conference and Exhibition o f the Society o f Petroleum Engineers SPE 28428.
Jenyon, et al. (1987). The Southern Zechstein salt basin o f the British North
Sea, as observed in regional seismic traverses. Petroleum Geology o f North W est
Europe: Proceedings o f the 3rd Conference on Petroleum Geology o f North W est
Europe. J. Brooks and K. W. Glennie. London, Graham & Trotman. 1: 277-292.
Katahara, K. W. (1995). “Gamma ray log response in shaly sands.” The Log
Analyst 36(4): 50-56.
Kearey, P., et al. (1991). An introduction to geophysical exploration. Oxford,
Blackman Scientific Publications.
Kenyon, W. E. (1997). “Petrophysical principles o f applications o f N M R
Logging.” The Log AnalvstfMarch-Aprill: 21-43.
Khaksar, A., et al. (1998). “Porosity from sonic log in gas-bearing shaly
sandstones: field data versus empirical equations.” Exploration Geophysics 29: 440-
446.

B-7
 Bibliography

Khaksar, A., et al. (1999). “Influence o f effective stress on the acoustic

velocity and log-derived porosity.” SPE Reservoir Evaluation and Engineering 2(1):
69-74.
Kimball, C. V., et al. (1998). “Error bars for sonic slowness measurements.”
Geophysics 63(21: 345-353.
Knox, R. W. O. B., et al. (1981). Stratigraphical relationships o f the
Palaeocene sands in the UK sector o f the central North Sea. Petroleum Geology o f the
Continental shelf o f north-west Europe. L. V. Illing and G. D. Hobson. London, The
Institute o f Petroleum: 267-281.
Koelman, J. M. V. A., et al. (1996). Interpretation o f resistivity logs in
horizontal wells: an application to complex reservoirs from Oman. 37th Annual
Logging Symposium Transactions, Society o f Professional Well Log Analysts.
Kulpecz, A. A., et al. (1990). Geological Modeling o f a Turbidite Reservoir,
Forties Field, North Sea. Sandstone Petroleum Reservoirs. M. a. S. Barwis. New
York, Springer-Verlag New York Inc.: 489-507.
Leurer, K. C. (1997). “Attenuation in fine-grained marine sediments:
Extension o f the Biot-Stoll model by the "effective grain model".” Geophysics 62:
1465-1479.
Li, P., et al. (1994). On induction response in extended-reach and horizontal
well and its shoulder bed effect correction (abstract onlvV 2nd International Well
Logging Symposium Transactions, Beijing Chapter, Society o f Professional Well Log
Analysts.
Liu, X., et al. (1993). “Quantitative determination o f Tertiary sediment supply
to the North Sea basin.” AAPG Bulletin 81(9): 1482-1509.
Lofts, J., et al. (1995). “Reservoir characterization from downhole
mineralogy.” Marine and Petroleum Geology 12(3): 233-246.
Lovell, J. P. B. (1990). Cenozoic. Introduction to the Petroleum Geology o f
the North Sea. K. W. Glennie. London, Balckwell Scientific Publications. 1: 273-293.
LPS, L. P. S. (1994). “Topical Conference on Quality.” The Log
Analvstf September-Octoher): 13-70.
Maeso, C., et al. (1999). Formation evaluation from logging-while-drilling
data in a 6.5inch horizontal well - A North Sea case study. 40th Annual Logging
Symposium Transactions, Society o f Professional Well Log Analysts.

B-8
 Bibliography

Minette, D. (1996). Developments in understanding the physical foundations

o f formation density and lithologv logging. 37th Annual Logging Symposium
Transactions, Society o f Professional Well Log Analysts.
Minette, D. C., et al. (1995). “The accuracy and precision o f FEMWD density
tools.” SPWLA 36th Annual Logging Symposium Transactions 1: B1-B12.
Mitchell, A. H. G., et al. (1986). Sedimentation and Tectonics. Sedimentary
Environments and Facies. H. G. Reading. Oxford, Blackwell Science: 471-519.
Moake, G. (1991). A new approach to determining Compensated Density and
Pe values with a Spectral-Densitv Tool. 32nd Annual Logging Symposium
Transactions, Society o f Professional Well Log Analysts.
Moake, G., et al. (1996). Reduction o f standoff effects on LWD density and
neutron measurements. 37th Annual Logging Symposium Transactions, Society o f
Professional Well Log Analysts.
Morrison, P., et al. (1953). Experimental Nuclear Phvsics. New York, John
W iley & Sons Inc.
Morton, A. C., et al. (1993). Stratigraphic evolution o f sand provenance during
Paleocene deposition in the northern North Sea. Petroleum Geology o f Northwest
Europe: Proceedings o f the 4th Conference. J. R. Parker. London, The Geological
Society: 73-84.
Mound, D. G., et al. (1991). The Cyrus Field, Block 16/28, UK North Sea.
United Kingdom Oil and Gas Fields 25 Years Commemorative Volume. I. L. Abbotts.
London, The Geological Society: 295-300.
Myers, G. D. (1992). “A Review o f Nuclear Logging.” The Log Analvst/Mav-
June): 228-238.
Nieto, J. A., et al. (1995). Removal o f borehole induced noise from well logs.
36th Annual Logging Symposium Transactions, Society o f Professional Well Log
Analysts.
Norve, K. H., et al. (1989). “Field experience using the full suite MWD-
combination for reservoir logging and evaluation.” SPWLA 30th Annual Logging
Symposium Transcations 1: A1-A26.
Nurmi, R. (1995). “Horizontal Highlights.” Schlumberger Middle East Well
Evaluation Review 16: 19.
Paillard, D., et al. (1996). “Macintosh program performs time-series analysis.”
Eos Transaction o f the AGU 77: 379.

B-9
 Bibliography

Paske, W. C., et al. (1990). “Theory and implementation o f a borehole caliper

measurement while drilling.” 65th Annual Technical Conference and Exhibition o f
the Society o f Petroleum Engineers SP E 20562.
Patchett, J. G., et al. (1994). Inverse modeling using full nuclear response
functions including invasion effects plus resistivity. 35th Annual Logging Symposium
Transactions, Society o f Professional W ell Log Analysts.
Patterson, D., et al. (1997). Integration o f cross-dipole acoustics for improved
formation evaluation. 38th Annual Logging Symposium Transactions, Society o f
Professional W ell Log Analysts.
Peeters, M., et al. (1999). Invasion in space and time. 40th Annual Logging
Symposium Transactions, Society o f Professional Well Log Analysts.
Pegrum, R. M., et al. (1990). Hydrocarbon plays in the northern North Sea.
Classic Petroleum Provinces. J. Brooks. London, The Geological Society. Geological
Society Special Publications 50: 441-470.
Pickett, G. R. (1963). “A coustic character logs and their application in
formation evaluation.” Journal o f Petroleum Technology 15: 659-667.
Pickett, G. R. (1973). “Pattern recognition as a means o f formation
evaluation.” The Log Analyst 14(4): 3-11.
Posamentier, H. W., et al. (1991). Seismic expression and recognition criteria
o f ancient submarine fans. Seism ic Facies and Sedimentary Processes o f submarine
fans and turbidite systems. P. W eim er and M. H. Link. London, Springer-Verlag:
197-222.
Prilliman, J., et al. (1995). “Steering and evaluation o f horizontal wells using
LWD.” Canadian Well Logging S ociety Journal 20: 18-26.
Raiga-Clemenceau, J., et al. (1988). “The concept o f acoustic formation factor
for more accurate porosity determination from sonic transit time data.” The Log
Analyst(January-Februaiy): 54-59.
Ranney, L. (1939). “The first horizontal oil well.” The Petroleum Engineer
X(9): 25-30.
Raymer, L. L., et al. (1980). An improved sonic transit time-to-porositv
transform. 21st Annual Logging Sym posium Transactions, Society o f Professional
W ell Log Analysts.
Reynolds, T. (1994). “Quantitative analysis o f submaine fans in the Tertiary o f
the North Sea basin.” Marine and Petroleum G eology 11(2): 202-207.

B-10
 Bibliography

Rider, M. H. (1996). The Geological Interpretation o f Well Logs. Caithness,

Whittles Publishing.
Rochow, K. A. (1981). Seismic Stratigraphy o f the North Sea 'Palaeocene'
Deposits. Petroleum Geology o f the Continental Shelf o f North-West Europe.
London, Institute o f Petroleum. 1: 255-266.
Sakurai, S., et al. (1992). “Wireline - MWD Logoff for Formation Evaluation,
Kuparuk River Unit, Alaska.” The Log Analvstf September-QctobeU: 451-460.
Samworth, J. R. (1980). “Slimline dual detector density logging - A
semitheoretical but practical approach to correction and compensation.” Journal o f
Petroleum Technology!August!: 1316-1322.
Samworth, J. R. (1991). A lgorithms for compensated neutron logging - 57
Varieties. Fourteen European Formation Evaluation Symposium, London, SPWLA.
Samworth, J. R. (2000). Horizontal well log quality analysis.
Samworth, J. R., et al. (2000). “CALLISTO - Calibration (at Leicester) and in-
situ optimization - a new world-standard facility for the calibration o f nuclear well
logs.” Dialogue 1(3).
Samworth, R. (1998). Resistivity profiles in horizontal wells and the
possibility o f similar bedding effects on porosity tools.
Schlumberger (1989). Log Interpretation Principles and Applications.
Houston, Schlumberger Education Services.
Schlumberger (1995). Logging While Drilling. Houston, Schlumberger
Education Services.
Schlumberger (1999). Log Interpretation Charts. Houston, Schlumberger
Education Services.
Schlumberger (1999). Web Site (www.connect.slb.com).
Schroeder, T. S., et al. (1991). “Integration o f Measurement While Drilling
and Wireline Formation Evaluation data.” Fourteenth European Formation Evaluation
Symposium 1: P1-P15.
Scott, H. D., et al. (1982). “Dual porosity CNL count rate processing.”
£PE(11146): 16.
Scott, H. D., et al. (1994). Response o f a Multidetector Pulsed Neutron
Porosity Tool. 35th Annual Logging Symposium Transactions, Society of
Professional Well Log Analysts.
Serra, O. (1984). Fundamentals fo Well Logging. New York, Elsevier.

B -ll
 Bibliography

Sheriff, R. E., et al. (1995). Exploration Seismology. Cambridge, Cambridge

University Press.
Sherman, H., et al. (1975). “The Depth o f investigation o f neutron and density
sondes.” SPWLA 16th Annual Logging Symposium Transactions 1: Q1-Q14.
Sherman, H., et al. (1983). Effect o f oil on the CNL. 24th Annual Logging
Symposium Transactions, Society o f Professional Well Log Analysts.
Singer, J. M. (1992). “An Example o f Log Interpretation in Horizontal Wells.”
The Log AnalvstfMarch-April!: 85-95.
Society, L. P. (2001). “Petrophysical Challenge: Spot the clay point.”
Dialogue 9(2).
Spross, R. L., et al. (1995). “Accurate MWD Density Measurements with
Very Large Standoffs.” SPWLA 36th Annual Logging Symposium Transactions 1:
UU1-UU12.
Staub, H., et al. (1953). Experimental Nuclear Physics. New York, John Wiley
& Sons Inc.
Stewart, I. J. (1987). A revised stratigraphic interpretation o f the Early
Palaeogene o f the central North Sea. Petroleum Geology o f North West Europe:
Proceedings o f the 3rd Conference on Petroleum Geology o f North West Europe. J.
Brooks and K. W. Glennie. London, Graham & Trotman. 1: 557-576.
Stow, D. A. V. (1986). Deep Clastic Seas. Sedimentary Environments and
Facies. H. G. Reading. Oxford, Blackwell Science: 399-444.
Tang, X. (1999). “Simultaneous inversion o f formation shear-wave anisotropy
parameter from cross-dipole acoustic-array waveform data.” Geophysics 64: 1502-
1511.
Taylor, J. C. M. (1990). Upper Permian - Zechstein. Introduction to the
Petroleum Geology o f the North Sea. K. W. Glennie. London, Blackwell Scientific
Publications: 153-190.
Theys, P., et al. (1996). An MWD Glossary. IMS.
Tiab, D., et al. (1996). Petrophvsics: Theory and practice o f measuring
reservoir rock and fluid transport properties. Houston, Gulf Publishing Company.
Tittle, C. W. (1961). “Theory o f neutron logging I.” Geophysics 26: 27-39.
Tittle, C. W., et al. (1966). “Theory o f neutron logging II.” Geophysics 31:
214-224.
Tittman, J. (1986). Geophysical Well Logging. Acadmic Press Inc.

B-12
 Bibliography

Tittman, J., et al. (1965). “The physical foundations o f formation density

logging (gamma-gamma).” Geophysics XXX(2): 284-294.
Vernik, L., et al. (1997). “Velocity anisotropy in shales: A petrophysical
study.” Geophysics 62(2): 521-532.
Vernon, G. K., et al. (1993). Successful applications o f horizontal-well
induction modeling. 34th Annual Logging Symposium Transactions, Society o f
Professional Well Log Analysts .
Wahl, J. S., et al. (1964). “The Dual Spacing Formation Density Log.” Journal
o f Petroleum Technology 16: 1411.
Walker, I. M., et al. (1987). The structural and stratigraphic evolution o f the
northeast margin o f the Sole Pit Basin. Petroleum Geology o f North West Europe:
Proceedings o f the 3rd Conference on Petroleum Geology o f North West Europe. J.
Brooks and K. W. Glennie. London, Graham & Trotman. 1: 263-275.
Weber, K. J. (1999). “The Prize-what's possible?” Petroleum Geoscience 5(2):
135-144.
Went, D. J., et al. (1997). “Integration o f horizontal well geological data into
reservoir description: Rotliegende Sandstone, Lancelot area, southern North Sea.”
AAPG Bulletin 81(1): 135-154.
Western, et al. (1985). Log Interpretation Charts. Atlas Wireline Services.
White, J. (1991). “Recent North Sea experience in formation evaluation o f
horizontal wells.” SPE(23114): 12.
Whitman, W. W., et al. (1992). “The influence o f elastic and density
properties on the behaviour o f the Gassmann relation.” The Log Analyst(Novemher-
December): 500-506.
Whyatt, M., et al. (1992). The Nelson Field: a successful application o f a
development geoseismic model in North Sea exploration. Exploration Britain. R. F. P.
Hardman. London, The Geological Society. The Geological Society Special
Publication No. 67: 283-305.
Wiley, R., et al. (1994). The effects o f invasion on density/thermal neutron
porosity interpretation. 35th Annual Logging Symposium Transactions, Society o f
Professional Well Log Analysts.
Winsauer, W. A., et al. (1952). “Resistivity o f brine-saturated sands in relation
to pore geometry.” Bulletin o f the American Association o f Petroleum Geologists
36(2): 253-277.

B-13
 Bibliography

Wolcott, D. S., et al. (1995). Characterization o f a heterogenous formation

using azimuthal lo g ging-while-drilling (LWD1 measurements. 36th Annual Logging
Symposium Transactions, Society o f Professional Well Log Analysts.
Wood, A . W . (1941). A textbook o f sound. New York, The Macmillam
Company.
Woodhouse, R ., et al. (1991). Vertical migration o f invaded fluids in
horizontal w ells. 32nd Annual Logging Symposium Transactions, Society o f
Professional W ell L og A nalysts.
Worthington, P. F. (1985). “The evolution o f shaly-sand concepts in reservoir
evaluation.” The L og A nalyst 26(1): 23-40.
Wraight, P., et al. (1989). “Combination Formation Density and Neutron
Porosity measurements w hile drilling.” SPWLA 30th Annual Logging Symposium
Transcations 1: B 1-B 23.
Wu, J.-Q., et al. (1997). Measurement o f dip angle and horizontal and vertical
resistivities using m ultiple frequency propagation resistivity tools. 38th Annual
Logging Symposium Transactions, Society o f Professional Well Log Analysts.
W yllie, M. R. J., et al. (1956). “Elastic wave velocities in heterogeneous and
porous media.” G eophysics 2: 41-70.
Wyllie, M. R. J., et al. (1956). “Elastic wave velocities in heterogeneous and
porous media.” G eophysics 2: 41-70.
Xu, S., et al. (1995). “A new velocity model for clay-sand mixtures.”
Geophysical Prospecting 43: 91-118.
Ziegler, W. H. (1975). Outline o f the Geological History o f the North Sea.
Petroleum and the Continental shelf o f North-West Europe. W. A. W. Barking,
Elsevier Applied S cien ce Publishers: 165-189.

B-14