Pore Pressure Trend Analysis

2007 Weatherford International Ltd.

Training Curriculum
CRCM_390_revB_0704

T R AI N
& E T. 200 I N
S 0
PD

E
G
ES
EMP

YE

W
O

ER L
ING EMP
This page intentionally left blank.
Formation Pore Pressure
Trend Analysis

Basic Concepts

• Pressure
– The force that a fluid (liquid or gas) exerts uniformly
in all directions within a vessel, pipe, wellbore,
formation, etc. (psi)
• Pressure Control
– The act of preventing the entry of formation fluids
into a wellbore
• Pressure Differential
– The difference between two fluid pressures; for
example, the difference between pressure in a
reservoir and in a wellbore drilled in the reservoir

1
Basic Concepts

• Formation Pore Pressure

– The force exerted by fluids in a formation
• Overburden Pressure
– The pressure exerted by the rock strata on a formation
of interest
– Generally considered to be approximately 1.0 psi/ft
• Fracture Pressure
– The pressure at which a formation will break down

Basic Concepts

• Pressure Gradient
– A scale of pressure differences in which there is a
uniform variation of pressure from point to point
– The normal pressure gradient in a formation is
equivalent to the pressure exerted at any given depth
by a column of 10% salt water extending from that
depth to the surface (0.465 psi/ft)
– The pressure gradient of pure water is approximately
0.433 psi/ft
• Fracture Gradient
– The pressure gradient at which a formation accepts
whole fluid from the wellbore

2
Basic Concepts

• Mud Weight (MW)

– A measure of the density of a drilling fluid in pounds
per gallon (ppg)
• True Vertical Depth (TVD)
– The measured depth of the well projected into the
vertical plane
• Hydrostatic Pressure (HSP)
– the amount of pressure the column of drilling mud
exerts at the bottom of the hole
– HSP = 0.052 x MW x TVD

Hydrostatic Pressure

10 ppg mud

TVD

HSP = HSP = HSP

3
Derivation of the HSP Equation

• Consider a column of fluid:

– Cross-sectional area = 1 sq.ft.
– Height = TVD ft
– Density = MW
• Weight of the fluid = Volume*Density
=Area (ft2)*TVD (ft)*62.4 lb/ft3*MW/8.33 lb/gal
=7.491*MW*TVD

Derivation of the HSP Equation

• This weight is equally distributed over an

area of 1 ft2 (144 in2)
• Weight (force)/area
= 7.491*MW*TVD/(144 in2)
• HSP = 0.052*MW*TVD

4
Exercise #1

• Calculate the HSP for each of the following:

– 10,000 ft of 12.0 ppg mud
– 12,000 ft of 10.5 ppg mud
– 15,000 ft of 15.0 ppg mud

Exercise #1 Solution

• HSP = 0.052*MW*TVD

– 0.052*12*10,000 = 6240 psi

– 0.052*10.5*12,000 = 6552 psi

– 0.052*15.0*15,000 = 11,700 psi

5
Exercise #2

• For a hydrostatic pressure equal to 5,000 psi,

what are the calculated vertical depths for
each of the following mud weights:
– 8.33 ppg
– 10 ppg
– 16.5 ppg

Exercise #2 Solution

• TVD = HSP / (0.052*MW)

– 5,000/(0.052*8.33) = 11,543 ft

– 5,000/(0.052*10) = 9615 ft

– 5,000/(0.052*16.5) = 5828 ft

6
Exercise #3

• What mud weights (to 2 decimal places) are

necessary to balance a formation pressure of
5000 psi at the following true vertical depths:
– 8000’
– 9,500’
– 11,000’

Exercise #3 Solution

• MW = HSP / (0.052*TVD)

– 5,000/(0.052*8000) = 12.02 ppg

– 5,000/(0.052*9,500) = 10.12 ppg

– 5,000/(0.052*11,000) = 8.74 ppg

7
Equivalent Circulating Density

• Equivalent Circulating Density (ECD)

– The increase in bottomhole pressure expressed as an
increase in pressure that occurs only when mud is
being circulated
– Because of friction in the annulus as the mud is
pumped, the bottomhole pressure is slightly higher
than mud that is not being pumped

– ECD = Mud Weight + Annular Pressure Loss

0.052 x TVD

Exercise #4

• What is the ECD at 5,000’ true vertical depth

with a mud weight of 10 ppg in the hole if the
annular pressure losses are:
– 50 psi
– 75 psi
– 125 psi

8
Exercise #4 Solution

• ECD = Mud Weight + Annular Pressure Loss

0.052 x TVD

– 10 + (50/(0.052*5000)) = 10.19 ppg

– 10 + (75/(0.052*5000)) = 10.29 ppg

– 10 + (125/(0.052*5000)) = 10.48 ppg

Compaction Theory
• During deposition, sediments are compacted
by the overburden load and are subjected to
greater temperatures with increasing burial
depth
• Porosity is reduced as water is forced out

9
 Compaction Theory

• Hydrostatic equilibrium within the

compacted layers is retained as long as the
expelled water is free to escape
• If water cannot escape, abnormal pressures
occur

Compaction Theory
Pore water expelled
Undercompacted
because of increasing
Shales
overburden

If the expelled water is not free

to escape, abnormal pressures
may result. Sufficient
compaction cannot occur so the
pore fluids carry more of the
overburden

10
Normal Pressure

• Normal subsurface
fluid pressure
distribution in
shallow deltaic
sediments

Compaction Theory

σob = σeV + pp
σob = overburden stress
σeV = vertical matrix
stress
pp = pore pressure

The overburden load is supported by the

vertical stress in the grain framework
and by the fluid pore pressure

11
Compaction Theory
• The average porosity in
sediments generally
decreases with increasing
depth due to the increasing
overburden
• This results in an
increasing bulk density
with increasing depth, and
increasing rock strength

Pressure Gradients

Normal pressure
gradients
correspond to the
hydrostatic gradient
of a fresh or saline
water column

12
Normal Pore Pressure Gradients

Pore Pressure

Pore pressures do not

always increase with
depth however

13
Subnormal Pressures
• Formation pressure gradients less than
normal gradients for a given area

• Lost circulation problems and differential

sticking are common problems in these
areas

Faulting

8000’
9000’

14
Aquifer outcrops below rig

Production of oil or gas

15
Abnormal Pressure Facts
• Abnormal Pressures are formation pressures
greater than normal pressures
• Many abnormal pore pressure processes are
simply the reverse of those which effect
subnormal pressures
• All abnormal pressures require some means
of sealing or trapping the pressure within the
rock body, otherwise hydrostatic equilibrium
back to a normal gradient would eventually
be restored

Abnormal Pressure
• Massive shales provide good pressure seals,
but shales do have some permeability, so,
given sufficient time, normal pressures will
eventually be established
• It may take tens of millions of years for a
normal pressure gradient to re-occur
• Dense rocks should always be a warning to
a driller that the pore pressure may be
changing

16
Aquifer

Thick gas sand

2 3
P = 605 - 0.05 * 300 g = 590/1000
= 605 - 15 = 0.590 psi/ft
= 590 psig EMW = 0.590/0.052
1 11.4 ppg
p = 0.465 * 1300
= 605 psig

17
Reverse Fault

9000’
10,000’

Normal Fault
Top of
Transition
Zone

Pressure may
increase due to
flow across fault
boundary

18
Salt Diapirs

Salt diapirs
plastically “flow” or
extrude into the
previously deposited
sediment layers
The resulting
compression can
create overpressure.

Erosion

19
Caprock Mineral Deposition

Man-Made Abnormal
Pressures

Underground Casing Faulty

blowout leaks cement job

20
 Fracture
Pressure
Line

Pore
Pressure
Line
Sa
fe

Lost Returns
Op

Area
era
tin
Wg
ind
ow

Kick
Area

21
What is a Kick?
• “Kick”
– An unscheduled entry of formation fluids into
the wellbore of sufficient quantity to require
shutting in the well
• What is a Blowout?
– Loss of control of a kick

Why does a kick occur?

• Pressure in the wellbore is less than the

pressure in the formation
• Permeability of the formation is great
enough to allow flow

22
How do we prevent kicks?
• We must maintain the pressure in the
wellbore greater than formation pressure
• However, we must not allow the pressure in
the wellbore to exceed the fracture pressure
• This is done by controlling the HSP of the
drilling fluid, and isolating weak formations
with casing

What is Lost Circulation?

• “Lost” Circulation
– Quantities of whole mud lost to a formation,
usually in cavernous, fractured, or highly
permeable beds
– Lost circulation can lead to a blowout, and in
general, can reduce the efficiency of the drilling
operation
– Commonly called “lost returns”

23
Why does lost circulation occur?
• Typically because the pressure in the
wellbore is greater than the fracture
pressure of the formation
• Evidenced by the complete or partial failure
of the mud to return to the surface as it is
being circulated in the hole

How do we prevent lost

circulation?

• We must maintain the pressure in the

wellbore below formation fracture pressure
• This is done by controlling the HSP and
ECD of the drilling fluid, and isolating
weak formations with casing

24
Leak Off Test

Bad Leak Off Test

• Leak Off Test

results indicating
a poor cement
bond

25
Resistivity and Conductivity
• The ability of rock to conduct electric current
can be used to infer porosity
• Resistivity -- ohm-m2/m or ohm-m
• Conductivity -- 10-3m/ohm-m2 or millimhos/m
• Conductivity = 1000/Resistivity

Resistivity and Conductivity

• Rock grains, in general, are very poor

conductors
• Saline water in the pores conducts
electricity and this fact forms the basis for
inferring porosity from bulk resistivity or
conductivity measurements

26
Resistivity and Conductivity

• Under normal compaction, resistivity

increases with depth
• Deviation from the normal trend suggests
abnormal pressure

Resistivity and Conductivity

• FR = Ro/Rw • FR = formation
resistivity factor
• Ro = resistivity of
water saturated
formation
• Rw = resistivity of pore
water

27
Resistivity of formation water

• Rw reflects the dissolved salt content of the

water, and is dependent upon temperature
⎛ T + 6.77 ⎞
Rw 2 = Rw1 ⎜⎜ 1 ⎟⎟
⎝ T2 + 6.77 ⎠
Where T1 and T2 are in deg F
• Equation shows that Rw decreases with
increasing temperature, e.g. decreases with
depth

Porosity, φ

φ = aFR−1/ m
•If a = 1, and m = 2, then
φ = (Ro/Rw)-0.5
•Rw in shales cannot be measured directly so
Rw in a nearby sand is used instead
•Ro would tend to increase with increasing
depth under normally pressured conditions

28
“Normally”
Pressured Wells

• Most of the plotted

shale conductivities
fall along the
“normal” trend line
in each of these
wells

“Abnormally”
Pressured Wells

• The “transition”
from normally
pressured shale
to overpressured
shale is where
the points deviate
from the normal
line

29
Transition Zone

Top of Pressure Transition

Older shales have had more

time to compact, so porosities
would tend to be lower (at a
particular depth).

Use the trend line closest to

the transition.

Lines may or may not be

parallel.

30
 Overburden stress
depends upon porosity,
and porosity depends on
overburden stress
Shales are more
compactible than
sandstones.
Young shales are more
compactible than older
shales.
Limestones and dolomites
are only slightly
compactible.

Eaton’s Equations 3
⎛ ∆t n ⎞
g p = g ob − ( g ob − g n )⎜⎜ ⎟⎟
⎝ ∆to ⎠
1.2
⎛R ⎞
g p = g ob − ( g ob − g n )⎜⎜ o ⎟⎟
⎝ Rn ⎠
1.2
⎛C ⎞
g p = g ob − ( g ob − g n )⎜⎜ n ⎟⎟
⎝ Co ⎠
1.2
⎛d ⎞
g p = g ob − ( g ob − g n )⎜⎜ co ⎟⎟
⎝ d cn ⎠

31
Eaton’s Equations

• These equations differ from the earlier

correlations in that they take into
consideration the effect a variable
overburden stress may have on the effective
stress and the pore pressure.
• Probably the most widely used of the log-
derived methods
• Have been used over 20 years

32
 Prediction of
Abnormal Pore Pressure

Pore Pressure Prediction Methods

• Resistivity of Shale
• Temperature in the Return Mud
• Drilling Parameters
• Modified dc - Exponent
• Sonic Travel Time
• Shale Resistivity

1
Shale Resistivity
vs. Depth

1. Establish trend
line in normally
pressured shale

2. Look for
deviations from
this trend line

(semi-log)

EXAMPLE

Shale Resistivity
vs. Depth

1. Establish normal
trend line

2. Look for
deviations

(semi-log)

2
 Pore Pressure
Shale Resistivity (lb/gal equivalent)
vs. Depth 16 14 12 10

1. Establish normal
trend line
9 ppg
2. Look for
(normal)
deviations

3. Use OVERLAY
to quantify
pore pressure
(use with caution)
Depth, ft

Shale Density , g/cc

3
Depth, ft

Mud Temperature in flowline, deg F

Example

8.2 X

Why?

4
 Example

8.8 X

Thermal conductivity, heat capacity, pore pressure...

Drilling Rate, ft/min

PHYD - PPORE , psi

5
 ∆P = (P2 - P1)1,000

Effect of Differential Pressure

Typical Drilling Rate Profiles - Shale

Shale The drilling rate in a normally

pressured, solid shale
section will generally
generate a very steady and
smooth drilling rate curve.

The penetration rate will be

steady and not erratic
(normally pressured, clean
shale).

6
 Typical Drilling Rate Profiles - Sand

Sand The drilling rate in a sand will

probably generate an erratic
drilling rate curve.

Sands in the Gulf Coast area

are generally very
unconsolidated. This may
cause sloughing, accompanied
by erratic torque, and
temporarily, erratic drilling
rates.

Typical Drilling Rate Profiles - Shaley Sands

This is generally the most

Shaley Sands troublesome type drilling rate curve
to interpret.

Many times this curve will look

similar to a solid shale curve that
is moving into a transition zone.

Note: This is a prime example why you should not base

your decision on only one drilling parameter, even
though the drilling rate parameter is one of the better
parameters.

7
 Typical Drilling Rate Profiles

Transition Zone
Shale If you are drilling close to
balanced, there will probably
be a very smooth, (gradual)
increase in the drilling rate.

This is due to the difference

between the hydrostatic
head and the pore pressure
becoming smaller. ∆p

Typical Drilling Rate Profiles

As the pressure ∆p becomes
Transition Zone very small, the gas in the pores
Shale has a tendency to expand which
causes the shale particles to pop
from the wall. This is called
sloughing shale.
The transition zone generally
has a higher porosity, making
drilling rates higher. In a clean
shale the ROP will increase in a
smooth manner.

8
 Typical Drilling Rate Profiles

Note:

• If you are drilling overbalanced in a transition it

will be very difficult to pick up the
transition zone initially.

• This will allow you to move well into the

transition zone before detecting the problem.

Typical Drilling Rate Profiles

• This could cause you to move into a permeable

zone which would probably result in a kick.

• The conditions you create with overbalanced

hydrostatic head will so disguise the pending
danger that you may not notice the small
effect of the drilling rate curve change. This
will allow you to move well into that transition
zone without realizing it.

9
This page intentionally left blank.
 Pore Pressure Detection

1) Over Pressure Problems

a) Hole enlargement / collapse
b) Loss of circulation
c) Stuck pipe
d) Kicks
e) Blowouts
f) Inadvertent fracturing of formation
2) Down Hole Pressure
a) Overburden pressure: due to the weight of the above sediment.
b) Matrix stress pressure: portion of the overburden supported by the solid rock.
c) Pore pressure
d) Fracture pressure
e) Drilling fluid pressure
3) Factors Affecting Pore Pressure
a) Surface communication
b) Concentration of dissolved salts
c) Precentage of gas
d) Magnitude of existing temperature gradients
e) Precentage of effective porosity
f) Magnitude of overburden
4) Pore Pressure
a) Normal pore pressure: Gulf Coast .465 psift
b) Subnormal: lower than normal
c) Abnormal: higher than normal

Techniques for Pore Pressure Detection

1) Before Drilling
a) Seismic velocity profile
b) Offset well data
2) During Drilling
a) Drill rate
b) MWD gamma and resistivity logging
c) Measurement, connection and trip gas
d) Examination of drill cuttings
e) Shale density will began to decrease as pore pressure increases
f) Torque, drag, fill in the hole
g) Drilling mud salenity, viscosity, and temperture increases
3) After Drilling
a) wirelog
b) Direct pressure measurement
 Pore Pressure Calculations

Eaton’s Method Normal Trend Line

P = S - [( S - Pn ) x Ro ]^ 1.2 Rn = Ro ( S - P ) ^ -.833
Rn S - Pn

Ro = Rn ( S - P ) ^.833
S - Pn
P= pore pressure
S= overburden
Pn= normal pore pressure (GC= .865)
Ro= shale resistivity
Rn= normal resistivity (normal trend line)

Leak Off Formula Fracture Pressure Formula

F= MW + Leakoff Pressure F= V * (S - P) + P
(0.052)TVD (1 - V)
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

OVERBURDEN PRESSURE

Introduction

The calculation of overburden pressure and the construction of an overburden pressure

(or gradient) profile, or equation, is of prime significance to good well planning.
Overburden pressure is an essential input to quantitative pore pressure techniques.
Although rough quantitative estimations of pore pressure can be made through
observance of the relationship between mud weight and connection/trip gas, and by use
of locally-developed graphical overlays, few users would dispute the fact that accurate
estimations of pore pressure cannot be made without an accurate estimation of
overburden pressure. In addition, overburden pressure is an essential input for estimation
of fracture pressure. Hence, good well planning and execution cannot be achieved
without the determination of overburden pressure.

Definition of, and Factors Influencing, Overburden Pressure

Overburden pressure is defined as a pressure acting vertically downwards from the

earth’s surface, the magnitude of which is dependent upon the densities of the sediment
matrices and the densities of the sediment pore fluids. Overburden pressure is referenced
to a plane parallel to the earth’s surface and at a chosen true vertical depth. For land
locations, overburden pressure begins at the land surface; for offshore locations,
overburden pressure is referenced from the drill floor, and therefore incorporates the air
gap between the drill floor and mean sea level, the depth of the sea, and the vertical
thickness of underlying sediments.

Many years ago, numerous quantitative pore and fracture pressure methods made the
assumption that the overburden gradient was a constant, equal to 1.0 psi/ft. This
assumption would require the density of all formations below the surface reference plane
to be equal to 2.30 g/cc. The assumption led to frequently catastrophic errors in
estimations of pore and fracture pressures, especially for offshore locations.

The expression for overburden pressure is:

S = 0.433 ∫0Z ρ(z)dz

where:
S = overburden pressure, psi
ρz = sediment bulk density at depth z, g/cc
dz = thickness of interval, ft

1
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Definition of, and Factors Influencing, Overburden Pressure

The above equation yields overburden pressure, in pressure units (psi in this case) only
for an interval of thickness z. By accumulating the overburden pressure of all intervals,
the cumulative overburden pressure is determined at any given (vertical) depth.

The range in typical sediment densities is from approximately 1.8 to 3.15 g/cc, equivalent
to gradients ranging from 0.78 to 1.36 psi/ft. Depositional basin water densities range
from 1.0 to 1.08 g/cc, producing overburden gradients ranging from 0.433 to 0.468 psi/ft.
Consequently, overburden gradients are variable. The natural variety of formation and
water densities suggests that each drilling location will exhibit a unique overburden
pressure profile. Having said this, the differences in overburden gradient profiles
between different locations may range from negligible to significant.

Whereas the differences in overburden gradients between onshore locations may be

small, the difference between onshore and offshore and between offshore locations in
differing depths of water may be considerable. Offshore water depths have a significant
impact on the magnitudes of the overburden pressures. The nature of this effect, for
example, is a reduction in fracture pressures. Hence, water depth is a powerful
determinator in offshore casing setting depths.

For offshore overburden pressures, one of the contributing pressure intervals is composed
of seawater, possessing a much lower density than sediment. As a consequence, the
interval (depth) of water contributes a much lower interval overburden pressure than an
equivalent thickness of sediment. The deeper the water, the lower the overburden
pressure at a given depth as compared to a location of shallower water depth.

Another influence on overburden pressures unique to offshore situations is the air gap.
The air gap is the distance from the drill floor (representing the datum from which most
depths and pressures are measured) to the water line. For offshore rigs this distance
typically ranges from 60 to 80 feet, but may exceed 100 feet. Although this interval of
air may seem insignificant, its impact may have as much as a 1.0 ppg influence on
decreasing the cumulative overburden pressure, particularly at shallow sediment
penetration depths.

2
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Sources of Overburden Pressure Information

The determination of overburden pressure requires determination of water and formation

densities and conversion of these densities to pressures. The sources of formation
density data are several:

1. Drilled cuttings;
2. Sonic transit time log;
3. Seismic travel time log;
4. Formation density log;
5. Offset well or regional data curves or equations.

With the exception of last of the above sources, each source measures a property that
differs in kind or degree, but nonetheless provides an approximation to formation density
and, ultimately, to overburden pressure. Regardless of the source of data, the method for
converting density to pressure is the same.

In terms of reliability and confidence in the data used to determine a local overburden
gradient profile, the above sources may be ranked in order of greatest to least:

1. Formation density log;

2. Sonic transit time log;
3. Offset wells or regional data curves or equations;
4. Seismic travel time log;
5. Drilled cuttings.

Although the formation density log provides the best source of overburden pressure data,
density logs are rarely available for shallow hole sections. This, then, makes the
sonic log the best source. However, a sonic log is not available until after drilling.
Therefore, offset well or regional overburden gradient curves or equations become,
by default, the most commonly used source of overburden gradient data. Although
the seismic travel time profile (similar to the sonic log) for the location is available
prior to the commencement of drilling, results (pressures) from the seismic profile
may be low if the seismic coverage (known as offset) is poor. Finally, drilled cuttings
may provide reasonable estimations, but measurement of cuttings densities is seldom
made because the techniques are either laborious or hazardous.

3
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

General Procedure for Construction of an Overburden Table

Density data are required in order to construct an overburden pressure table. The density
data may be immediately available from a bulk density log or from drilled cuttings, or
may be indirectly available from sonic transit time or seismic interval velocity logs.
Once density data are available, the procedure for converting densities into overburden
pressures is as follows.

1. Determine the average density for a particular interval.

It is usually acceptable, in terms of accuracy, to use intervals on the order of 250
or 500 feet in thickness.
2. Convert interval average bulk density (g/cc) to an interval pressure gradient (psi/ft).
ρb * 0.433 = psi/ft
3. Multiply interval pressure gradient (psi/ft) by interval thickness (ft) to determine an
interval overburden pressure (psi).
psi/ft * ft = psi
4. Sum the interval overburden pressures to produce a cumulative overburden pressure
(S, in psi).
psi1 + psi2 + …. + psiN = S
5. Divide cumulative overburden pressure (psi) by total true vertical depth for
cumulative overburden gradient (psi/ft); or convert to equivalent mud weight.

An example overburden table is shown below.

TVD, ft Interval, ft ρb, g/cc Gradient, Interval S, psi S, psi/ft S, ppg

psi/ft Pressure,
psi
0-100 100 0.00 0 0 0 N/A N/A
100-3360 3260 1.03 0.446 1454 1454 0.433 8.3
3360-3680 320 1.66 0.719 230 1684 0.457 8.8
3680-4200 520 2.00 0.866 450 2134 0.508 9.8

4
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Deriving Density from the Sonic Log

The method for deriving formation density from a sonic log (or seismic velocity profile)
is based on the principle that velocity of a sound wave through a medium is a function of
that medium’s density (it is also a function of that medium’s plasticity and Poisson’s
ratio). Since many oil company operators acquire sonic transit time data from surface to
well total depth, the sonic log often represents the best source for formation density data.

The method is described as follows:

1. Using a sonic log, choose an interval with a fairly consistent sonic transit time value.
The chosen interval may be a thin as 50 feet, and should not exceed 500 feet in most
instances. Once an interval thickness is chosen, it may be repeated in increments
from surface to total depth. Generally speaking, beds thinner than 50 feet may be
ignored without sacrificing accuracy.

2. Calculate formation density from sonic transit time:

ρ = 2.75 – 2.11 ( ∆t - ∆tm ) / (∆t + ∆tf)

where:

ρ = formation density, g/cc

∆t = interval sonic transit time, µsec/ft
∆tm = matrix transit time, µsec/ft
∆tf = pore fluid transit time, µsec/ft

3. Construct an overburden pressure table as previously described.

This process of deriving density from sonic transit time requires a number of
assumptions, two of which are the values for matrix transit time and pore fluid transit
time. The user must make a determination on the lithology and refer to a table of typical
matrix and fluid transit times.

Table of Matrix and Fluid Transit Times

Lithology ∆tm = matrix transit time, µsec/ft

Sandstone, unconsolidated 52.6
Sandstone, consolidated 58.8
Clay/shale 160 – 47.0
Limestones 47.5
Dolomites 43.5
Anhydrite 50.0

5
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Salt 66.7

Fluid ∆tf = pore fluid transit time, µsec/ft

Seawater 190-200
Fresh water 218

6
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Deriving Density from the Sonic Log (cont’d)

Since most overburden pressure tables are developed for offshore Tertiary basins made
up predominantly of shale, silt and sand, and because these sediments may possess a wide
range in matrix transit times, constant matrix transit time values may be assumed. The
following table shows these constants.

Table of Matrix and Fluid Transit Time Constants

Lithology or Fluid Matrix or Fluid Transit Time Constant,

µsec/ft
Sands 50.0
Silts 48.0
Clay/shales 47.0
Seawater 190

Generally speaking, although the use of these constants introduces error, the degree of
error is sufficiently low for the purposes of determination of overburden pressures. A
potentially larger source of error is the eyeballing of an interval transit time. Particularly
in an alternating sequence of sands, silts, and clays, the determination of interval transit
time may become a function of considerable subjective error. For this reason, it is
recommended that the user determine interval transit time from the integrated transit time
(ITT) parameter.

Integrated transit time is plotted in the middle track of sonic logs, and is shown as a tick
mark. Each tick mark represents an interval corresponding to 1000 sec/ft. By summing
the ITT values for a given interval thickness, an average interval transit time may be
determined:

∆t (µsec/ft) = # ITT * 1000 / interval thickness, ft

7
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Using Density from the Density Log

The density tool utilizes a radioactive source to bombard the formation with gamma rays.
The amount and relative energy of returned gamma rays is a measure of the density of
the formation. As previously stated, few oil company operators acquire density log data
over the complete length of a well. When available, the density log may be used in lieu
of the sonic log; owing to its shallower depth of investigation, density log data may not
be any better than sonic log data.

Formation bulk density values may be read directly from the density log. Intervals
should be chosen to reflect a consistent density value. The user should utilize the density
correction curve as a quality indicator; although the correction has been applied to the
density value, density values from log intervals with high corrections should be
considered suspect.

8
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Historical Development of Overburden Pressure Determination Techniques

Throughout the 1960’s, overburden gradient was considered constant, equal to 1.0 psi/ft.
One of the earliest known released variable overburden profiles was that of Bell’s in
1969. Eaton published a paper in 1972 on the influence of overburden pressure for
determination of pore pressure, emphasizing the importance of considering overburden
gradients as being variable. Eaton also promoted the use of density logs as the preferred
source of data. In 1979, Belloti et. al., of AGIP, published their paper on the use of the
sonic log for determining overburden gradients.

Since 1979, very little has been published related to techniques for determining
overburden gradients. From time to time, a local or regional overburden curve has been
included in the published literature. Some authors have published equations related to
other authors’ curves; some of these equations are useful, while others are simply crude.

9
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Bell

Bell, L.N., ”Pressures and Fracture Gradients,” Atlantic Richfield Drilling Manual, June
1969.

Bell (ARCO) was amongst the earliest researchers to obtain formation density data for
the construction of regional overburden curves. Although the specific sources of his
density data are unknown, Bell produced regional overburden curves for the Louisiana
Gulf Coast, Texas Gulf Coast, and the mid-continent. It is likely that his source data
were skewed to onshore, versus offshore, wells, although this does not present a problem
if his curves are properly used. Each of his curves exhibits a variable overburden
gradient. The Louisiana overburden curve shows lower values throughout the range of
depths than for the Texas curve; this difference is due to the higher clay to sand ratio on
the Louisiana coast.

Discussion:

Although the sources of Bell’s data are unknown, in the nearly thirty years since, very
few other authors have published regional overburden curves, and these do not
substantially differ from Bell’s Gulf Coast curves. When used in a proper manner (i.e.,
making adjustments for air gap and water depth), Bell’s curves still yield reasonable
results, in the Gulf of Mexico, in the absence of locally obtained data.

10
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Eaton

Eaton, B.A., “The Effect of Overburden Stress on Geopressure Prediction from Well
Logs,” Journal of Petroleum Technology, August 1972, pp. 929-934.

Eaton (Universal Consultants) argued that overburden stress gradients were functions of
burial depths in sedimentary basins similar to the Gulf of Mexico. He stated that
overburden gradient could be determined by any means where bulk density information
was measured from the surface. He said overburden stress was related to density by the
following equation:

S = ρave g D

where:

S = overburden gradient
ρave = average bulk density
g = acceleration due to gravity
D = depth (vertical)

Bulk density data were available from density logs. By use of a cumulative averaging
scheme the bulk density values are converted to pressure units. Graphs of overburden
gradient versus depth could then be constructed.

Discussion:

Eaton published one of the first papers concerning the fact that overburden gradient was a
variable function of burial depth. The assumption that overburden gradient was a
constant of 1.0 psi/ft lead to significant errors in estimation of pore and fracture
pressures.

Eaton advised the use of the bulk density log for determination of overburden gradients.
However, few operators acquired bulk density logs over the entire drilled interval of a
well. Unfortunately, Eaton did not provide any guidance for what to do when the density
log was not available.

11
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Belloti, DiLorenzo, Giacca

Belloti, P., Di Lorenzo, V., and Giacca, D., “Overburden Gradient Calculation from
Sonic Log,” Sixth European Symposium Transactions, Society of Professional Well Log
Analysts, London, March 6, 1979.

Belloti et. al. (AGIP) introduced a procedure to determining overburden pressure from
the sonic log. Their use of the sonic log, in preference to the density log, was due to the
greater overall application of sonic logs. Their relationships were based upon earlier
work by Willie on compacted and cemented formations, and by Maxwell on uncemented
formations. Willie related rock porosity for cemented sands and shales to sonic transit
time:

φ = ( ∆t - ∆tm ) / (∆tf - ∆tm)

where:

φ = porosity
∆t = formation transit time, µsec/ft
∆tm = matrix transit time, µsec/ft
∆tf = pore fluid transit time, µsec/ft

Maxwell related porosity of uncemented sands and shales to sonic transit time:

φ = 1.228∆t - ∆tm / ( ∆t + ∆tf )

Porosity is related to density thusly:

ρ = ρm (1- φ) + ρfφ

where:

ρ = formation density
ρm = matrix density
ρf = pore fluid density

By combining equations and assuming constants ρm = 2.75, ρf = 1.03, ∆tm = 47, and ∆tf
= 200, Belloti et. al. derived the following equation for compacted and cemented
formations:

ρ = 3.28 - ( ∆t / 89 )

12
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Belloti, DiLorenzo, Giacca (cont’d)

and for uncemented formations:

ρ = 2.75 – 2.11 ( ∆t - 53 ) / (∆t + 200)

Noting that these two equations yielded similar results at low transit times, they
concluded that the latter equation produced satisfactory results in most instances for all
types of formations.

Belloti et. al. compared the results from obtaining density from the sonic log using the
above equation, and from density from the density log, and determined the differences
were minimal. Lastly, they illustrated the effect of increasing water depth on lowering of
offshore overburden gradients.

Discussion:

This is an excellent contribution and a very worthwhile technique for deriving an

overburden gradient curve. The puzzling question is, why did so many other researchers
harp on the use of the density log. As these authors showed in their paper, and has been
shown in subsequent years, any differences between deriving overburden from a sonic
log as compared to a density log are minimal. And the sonic log is not only usually
available over the entire drilled interval, but the measurement itself is deeper into the
formation than the density measurement, the sonic log is easier to use, and the sonic log
provides pore pressure information.

13
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Simmons and Rau

Simmons, E.L., and Rau, W.E., “Predicting Deepwater Fracture Pressures: A Proposal,”
63rd SPE Technical Conference and Exhibition, Houston, TX, Oct 2-5, 1988.

Simmons and Rau (Chevron) were primarily concerned with deriving feasible and
accurate means for determining fracture gradients for offshore locations where water
depths exceeded 350 feet. They proposed two techniques:

1. A modification of Eaton’s 1969 method to account for the effect of a column

of water on the effective stress (rock matrix stress);
2. A method of bulk density log integration for accurate overburden pressure
estimation.

Simmons and Rau opined that regionally averaged data (e.g., regional overburden curves)
are, at best, marginally prone to errors on the magnitude of fifteen to twenty percent.
They argued that accurate area or block data relating to effective stress were absolutely
necessary. They recommended the straight-forward integration of density data from the
bulk density log.

In the absence of density log information, the authors recommended using Eaton’s
overburden curve, which they represented in equation form as:

Goba = 0.8511934 * e[(lnDeff – 6.206593)2/84.36084]

where:

Goba = average overburden stress gradient, psi/ft

Deff = effective depth, equal to half the water depth plus the sediment depth, ft
e = base of Naperian logarithms

Discussion:

Simmons and Rau’s recommendation to use the bulk density log for accurate
determination of overburden pressure assumed availability of density data from the
mudline to total depth. However, they neglected to point out that in but few cases are
these data available. Oil companies seldom run the density log in top hole, as these data
are expensive and usually desired for only potentially productive horizons. Also, the
density log is a relatively shallow investigation tool, and will produce erroneous results in

14
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Simmons and Rau (cont’d)

mud-altered formations and enlarged bore holes. Hence, Simmons and Rau’s
unsupported conclusion that the density log will yield a more accurate estimation of
overburden pressure can only be viewed as doubtful.

The authors’ contingency position for determining formation pressures in the absence of
overburden pressure derived from density log data is to use Eaton’s overburden curve
(shown in Eaton, B.A., “Fracture Gradient Prediction and its Application in Oilfield
Operations,” Journal of Petroleum Technology, Oct 1969, pp. 1353-1360). Although
admittedly subject to errors in the range of twenty percent, they attempted to reduce the
error by modifying Eaton’s technique to account for the depth of water. They account for
this effect by defining an effective depth, equal to the vertical thickness of sediment plus
one-half the water depth. After determining the overburden gradient of the sediment,
they add in the overburden due to the total depth of the water to arrive at an overburden
pressure at the depth of interest. The authors fail to explain why they account for the
effect of water in this way.

The authors provide a very arcane expression (equation) for Eaton’s overburden gradient,
employing the natural exponent and logarithm. The equation for Eaton’s curve is much
better and more elegantly expressed as a polynomial.

15
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Bryant

Bryant, T.M., “A Dual Shale Pore Pressure Detection Technique,” SPE/IADC Drilling
Conference, New Orleans, LA, Feb 28-March 3, 1989.

The author proposed a regional overburden gradient equation based on Bell’s Louisiana
Gulf Coast curve, modified for air gap, water depth, and, to some degree, location.

The author suggests that, in the absence of better local data, this relationship may be used
to ascertain overburden gradients in Tertiary basins. The author presented this
relationship in the form of a polynomial equation:

S/D = E + 2.64E-5(D) – 1.97E-9(D2) + 6.60E-14(D3) – 5.97E-19(D4)

where:

S/D = overburden gradient, psi/ft

D = true vertical depth of interest, less air gap and water depth, ft
E = overburden gradient offset at mudline, psi/ft

The true vertical depth of interest is the total true vertical depth of interest less the air gap
and offshore water depth. Although the author suggested an overburden offset of 0.85 is
adequate in the absence of local data, the values typically used are 0.85 for Texas Gulf
Coast, 0.80 for Louisiana Gulf Coast, and 0.825 for offshore deepwater.

Overburden pressure contributed by the sediments is equal to the product of the above-
derived overburden gradient and the total vertical thickness of the sediments at the depth
of interest. The overburden pressure contributed by the seawater is the product of the
water’s density gradient and its depth. The total overburden pressure at the depth of
interest is the sum of the overburden pressure of the sediment and the overburden
pressure of the seawater. The overburden gradient at the depth of interest is the total
overburden pressure divided by the true vertical depth as measured from the drill floor
(which includes the air gap).

Discussion:

The author’s equation for a regional Gulf of Mexico overburden curve yields reasonable
results when adjusted for air gap, water depth, and location. The adjustment for location
is represented by his overburden offset factor E. The author suggests a value of E = 0.85
produces acceptable results; this is particularly the case for locations along the Texas
Gulf Coast. The value of E = 0.80 yields improved results for the Louisiana Gulf Coast.
For offshore Gulf of Mexico deepwater locations, a value of E = 0.825 produces good
results.

16
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Bryant (cont’d)

For locations outside the Gulf, such as the North Sea, regional curves or curve equations
for these areas should be used in preference to the Gulf of Mexico curves.

17
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Eaton and Eaton

Eaton, B.A., and Eaton, T.L., “Fracture Gradient Prediction for the New Generation,”
World Oil, Oct 1997, pp. 93-100.

The Eatons reviewed earlier published work of Ben Eaton because, as they stated, the
Eaton method (for fracture gradient determination) is the most commonly used method,
but is frequently used incorrectly and has been incorrectly modified by some users.

The authors recommended the use of bulk density data in 500 foot intervals, and account
for both air gap and water depth. However, they recognized a major problem to this
recommendation is the lack of bulk density data. They suggested that seismic data can be
relied upon to provide the sediment density data, and promoted a relationship between
seismic velocity and bulk density attributed to Gardener (Gardner et. al., “Formation
velocity and density – The diagnostic basis for stratigraphic traps,” Geophysics, vol. 39,
1974):

ρb = 0.23V0.25

where:

ρb = sediment bulk density, g/cc

V = seismic interval velocity, ft/sec

The authors offered that, given bulk density or seismic data, there is no excuse to not
develop an overburden gradient curve for anywhere in the world.

Using deepwater Gulf of Mexico data published by Barker and Wood (Barker, J.W., and
Wood, T.D., “Estimating Shallow below Mudline Deepwater Gulf of Mexico Fracture
Gradients,” Houston AADE Chapter Annual Technical Forum, April 2-3, 1997), the
authors constructed a new overburden gradient curve for the Gulf of Mexico.

Discussion:

The authors suggested the use of seismic interval velocity in the absence of bulk density
data for determining overburden pressure. However, their recommendation is to use an
aging (1974) empirical correlation attributed to Gardner, for deriving density from
velocity. The authors do not question or critique the relationship or the accuracy of
resulting derivations. What the Eatons did not say in their paper was that Gardner’s

18
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Eaton and Eaton (cont’d)

relationship was developed with a data set of Paleocene shale densities. His equation was
an approximation only for these data, and does not necessarily apply to shales of other
geological ages.

Given that the Eatons promoted the use of seismic data, it seems odd that they did not
discuss using the AGIP technique.

19
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Regional Overburden Gradient Curves and Equations

It is frequently the case that quantitative pore and fracture pressure evaluations are
hampered by lack of current or offset well sonic and density logs necessary to calculate a
variable overburden pressure profile. As a temporary aid to assist in such situations,
regional overburden curves or equations of these curves may be used.

1. Bell presented two Gulf Coast curves, one for the Texas Gulf Coast, the other for the
Louisiana Gulf Coast. Polynomial expressions for these curves are:

South Texas Gulf Coast:

S/D = 0.85 + 2.64E-5(D) – 1.97E-9(D2) + 6.60E-14(D3) – 5.97E-19(D4)

Louisiana Gulf Coast:

S/D = 0.80 + 3.16E-5(D) – 2.58E-9(D2) + 1.13E-13(D3) – 1.82E-18(D4)

2. Data from Herring for a North Sea regional overburden curve:

North Sea:

S/D = 0.61 + 4.36E-5(D) – 2.64E-9(D2) + 7.56E-14(D3) – 7.37E-19(D4)

3. Bryant’s overburden equations as derived from Bell’s curves:

South Texas Gulf Coast:

S/D = 0.85 + 2.64E-5(D) – 1.97E-9(D2) + 6.60E-14(D3) – 5.97E-19(D4)

Louisiana Gulf Coast:

S/D = 0.80 + 2.64E-5(D) – 1.97E-9(D2) + 6.60E-14(D3) – 5.97E-19(D4)

Deepwater Gulf:

S/D = 0.825 + 2.64E-5(D) – 1.97E-9(D2) + 6.60E-14(D3) – 5.97E-19(D4)

20
DEEPWATER PRESSURE SERVICE OVERBURDEN PRESSURE

Regional Overburden Gradient Curves and Equations (cont’d)

4. Simmons and Rau’s equation for the Gulf of Mexico:

Gulf of Mexico:

S/D = 0.8511934 * e[(lnDeff – 6.206593)2/84.36084]

5. Amoco’s technique for approximating sediment density (from the Amoco/Chevron

Pore Pressure Applications course):

ρavg = 16.3 + [(D-W-A)/3125]0.6

where:

ρavg = average sediment bulk density from mudline to total depth, ppg
D = depth of interest, ft
W = water depth, ft
A = air gap, ft

Eaton presented a regional overburden curve for the Gulf of Mexico in his 1969 paper,
and more recently, in his 1997 paper. Equations for these curves are not available.
However, Eaton’s 1969 curve would entail a value of E = 0.84 in a lesser approximation
of Bell’s Louisiana curve. His 1997 curve falls lower than his 1969 curve, particularly at
depths less than seven thousand feet.

21
This page intentionally left blank.
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

FRACTURE PRESSURE

Introduction

The determination of fracture pressures is an integral tool in well planning. Knowledge

of the magnitudes of fracture pressures in a well assists in locating when and where to set
a string of protective casing. This knowledge allows the planner to develop a viable mud
program with mud weights that will not cause fracture initiation and loss of drilling fluid
to the formation. Furthermore, fracture pressures possess significant importance to the
well control program, influencing the ability to withstand a kick.

Definition

Fracture pressure is usually defined as an applied pressure which, when equal to or

exceeding the sum of the formation’s horizontal matrix stress and its pore pressure,
produces a structural deformation known as a fracture. In practice, however, several
different fracture pressures are considered in oilfield operations. These pressures are
frequently not distinguished or specifically identified, such that confusion and
misunderstanding often exists. This situation largely characterizes the history of fracture
gradient literature. The two fracture pressures that are normally considered are fracture
initiation pressure and fracture injection pressure.

Fracture initiation pressure is defined as an applied pressure equal to or slightly

exceeding the sum of the formation’s tensile strength, the least horizontal stress,
borehole hoop stresses, and the pore pressure. At this pressure, the formation ruptures,
producing a fracture.

Fracture injection pressure is an applied pressure equal to or slightly exceeding the least
horizontal stress and the pore pressure. This pressure is sufficient to hold open and
propagate (extend) a pre-existing fracture, and is, by definition, less than the fracture
initiation pressure.

Since fracture initiation pressure includes borehole hoop stresses and formation strength,
fracture initiation pressure may range from ten to more than fifty per cent higher than
fracture injection pressure. It is therefore important to distinguish between these two
fracture pressures.

1
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Derivation of Fracture Pressure Theory

An understanding of fracture pressure requires a basic review of rock and soil mechanics.
Deformation of rock is the result of applied stress(es). Three principal stresses,
distinguished by magnitude, are recognized in rock mechanics. These stresses are:

1. a major stress ( σ1 );
2. an intermediate stress ( σ2 );
3. a minor stress ( σ3 ).

Each of these principal stresses is perpendicular to the other. Deformation, such as

folding, faulting, or fracturing, occurs when the magnitude of the major principal stress
σ1 is several times or more the magnitude of the minor principal stress σ3. The resulting
deformation is parallel to the plane of the intermediate stress σ2, and perpendicular to the
σ1σ3 plane.

In geological areas characterized by normal faulting, the major principal stress is aligned
perpendicular to the earth’s surface, with the intermediate and minor principal stresses
mutually perpendicular and aligned parallel to the earth’s surface. When σ1 exceeds σ3
by several times or more, the resulting deformation is usually a set of conjugate fractures
or normal faults, each perpendicular to the σ1σ3 plane, at an angle of approximately sixty
degrees to the minor stress σ3. Areas characterized by thrust faulting have their principal
stress parallel to the earth’s surface with the minor stress perpendicular to the earth’s
surface. Areas that are described as being tectonically relaxed have their major principal
stress perpendicular to earth’s surface, and with the intermediate stress equal to the minor
stress.

The presence of a borehole describes a situation somewhat different to that described by

rock mechanics. The borehole creates a very local stress concentration field (consisting
of hoop stresses) that is superposed on the regional stress field. These hoop stresses may
range from negligible to nearly the same as the regional stresses. To create a fracture, a
pressure must be exerted that exceeds the hoop stresses, the formation’s tensile strength,
and the least horizontal stress (and, naturally, the formation pore pressure). In most
cases, the resulting fractures are nearly vertical, and will align with the strike of the
intermediate horizontal stress. In tectonically relaxed areas where σ2 = σ3 , there will be
no preferred orientation of the resulting fractures. In the case where pre-existing
fractures are penetrated by the borehole, when a pressure equal to or slightly exceeding
the least effective stress is applied, the fractures in a plane normal to the least effective
stress (i.e., parallel to the strike of the intermediate horizontal stress) will open.

2
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Derivation of Fracture Pressure Theory (cont’d)

The significance of the above discussion lies not in the orientation of the stress field, but
in the relationship between the magnitudes of the major and minor stresses. Unstable
conditions occur when σ1 exceeds σ3 by some amount. Again, in the case of creating a
fracture, the applied pressure must exceed the least horizontal stress σ3, the borehole
hoop stresses, the formation’s tensile strength, and the formation’s pore pressure. In the
case of extending pre-existing fractures, the applied pressure must exceed the least
horizontal stress σ3 and the formation’s pore pressure.

Most researchers concerned with the subject of fracture pressures consider that
formations penetrated during the drilling process deform elastically, as opposed to
plastically. For elastic deformation, a rock deforms (i.e., undergoes strain) upon
application of a stress; once the stress is removed, the rock returns to its original shape.
For plastic deformation, the rock does not return to its original shape upon removal of the
stress. Some researchers consider all rocks as behaving in an elastic manner. Many
researchers consider sands, shales, limestones, dolomites, and anhydrites as being elastic;
and clays and salts as behaving plastically. Hence, there exist several approaches to
fracture pressure prediction in terms of deformational behavior:

1. Rocks are elastic, and each lithology has its own unique Poisson’s ratio.
2. Some rocks are elastic, others are plastic.

If a vertical compressive stress is applied to a cylinder of rock, the cylinder will deform
as a function of the stress and the rock’s response to that stress. For elastic deformation,
the length of the cylinder will decrease (axial strain) and the diameter of the cylinder will
increase (transverse strain). The ratio of the transverse strain to the axial strain is termed
Poisson’s ratio, and is expressed as:

µ = ( ∆d/d )/( ∆l/l )

where:
µ = Poisson’s ratio
∆d = change in diameter
d = original diameter
∆l = change in length
l = original length

and where these strains are functions of specific rock properties. In actuality, Poisson’s
ratio is more complex than indicated by the above relationship; Poisson’s ratio is related
to the Lame constants applicable to elastic deformation (whereby if the stress is relieved,

3
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Derivation of Fracture Pressure Theory (cont’d)

the cylinder returns to its original shape). Poisson’s ratio is not applicable for visco-
elastic or plastic deformation. Some researchers involved in oilfield fracture gradient
prediction chose to consider deformation only as elastic; a few have considered both.

From his studies on soil mechanics, Terzaghi (1926) is credited with developing the
expression relating overburden, effective (also frequently termed matrix) stress, and pore
pressures; this relationship is sometimes referred to in the literature as the effective stress
law:

S=σ+p

where:
S = total vertical stress due to density of overlying sediments;
also termed the overburden pressure
σ = vertical effective stress due to the rock frame
p = pore pressure, considered the neutral stress

Terzaghi viewed the relationship between the three principal earth stresses as:

σx = σy = (ν/1-ν) σz

where:
σx = effective stress in the x horizontal plane
σy = effective stress in the y horizontal plane
σz = effective stress in the z vertical plane
ν = Poisson’s ratio

Terzaghi also recognized that additional tectonic stresses are often superimposed upon
the horizontal effective stresses, such that σx may be greater than σy :

σx = (ν/1-ν) σz + σT

where:
σT = tectonic stress in the x horizontal plane

Terzaghi found that the usual range in values for Poisson’s ratio was from 0.25 to 0.30.

4
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Derivation of Fracture Pressure Theory (cont’d)

Terzaghi arrived at a fracture pressure relationship, which is essentially given by the

following equation

F = (2ν/1-ν)S + (1-3ν/1-ν)P

where:

F = fracture pressure
ν = Poisson’s ratio
S = overburden pressure
P = pore pressure

Another researcher in this field, M. Biot, derived a similar yet different fracture pressure
relationship:

F = (2ν/1-ν)S + α(1-3ν/1-ν)P

where:

α = a porosity factor equal to 1-Cr/Cb, where Cr is the compressibility of

the solid rock, Cb is the compressibility of the porous rock skeleton.

Biot’s relationship indicates that when there is no porosity (Cr = Cb), the pore pressure
term disappears. Despite their differences, both Terzaghi and Biot recognized that
conditions necessary to initiate a fracture required a pressure exceeding the horizontal
stress:

F > 2 (ν/1-ν) σz

where (ν/1-ν) σz is some portion of the vertical effective stress, which in turn is equal to
the horizontal effective stress.

At a later date, further research indicated the practical range for Poisson’s ratio extended
from 0.25 to 0.50.

5
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Hydraulic Fracturing and Fracture Pressure Tests

A number of procedures exist whereby information concerning earth stresses, fracture

pressures, and maximum allowable mud densities may be determined. These procedures
include hydraulic fracturing, and leak-off tests (LOT). These two procedures typically
yield different information. Hydraulic fracturing is performed to produce or stimulate an
increase in production from a reservoir (normally sandstone, hence elastic deformation).
The pressure applied during hydraulic fracturing must rupture the formation, then hold
the fractures open while fluid and proppant is pumped into the propagating fractures.
Consequently, hydraulic fracturing yields both fracture rupture pressure and fracture
injection pressure.

A leak-off test is typically performed after setting casing to determine if the formation
just below the shoe (often considered the weakest point in the following open-hole
section) will hold either a pre-selected pressure or else the maximum anticipated mud
density. During the leak-off test, drilling fluid is slowly pumped into the shut-off
annulus, and the pressure is recorded as a function of the volume pumped. A plot of this
pressure versus volume (stress vs. strain) schedule shows a linear relationship. If the
purpose of the LOT is to test to a pre-determined pressure (typical for development
wells), the test typically is concluded while the pressure-volume relationship is linear. If
the purpose of the LOT is to test to a maximum equivalent mud density (typical of
wildcat wells), the test continues until the relationship becomes non-linear; the pressure
corresponding to this point is termed the leak-off pressure. This is the pressure at which
the hoop stresses around the borehole begin to yield. If the test were to continue (and
assuming no pre-existing fractures are present), the pressure would continue to build to a
maximum, after which the pressure drops off very suddenly. This latter, maximum
pressure is the rupture pressure. If the test were to continue, the pressure would level off,
such that regardless of the pump rate the pressure remains approximately the same. This
last pressure is the injection, or propagation, pressure. Typically, the injection pressure
for sands is approximately 10% lower than the leak-off pressure; for shales, the rupture
pressure may be significantly higher than the injection pressure.

Of the two procedures, only hydraulic fracturing determines the injection pressure, which
is also equal to the least horizontal stress. The leak-off test typically only tests to the leak
off pressure. This pressure may range from one to two times the magnitude of the least
horizontal pressure. What this means is that fracture gradients determined from leak-off
tests are higher than fracture gradients determined from hydraulic fracturing. This
emphasizes the importance of knowing exactly which pressure is used to determine the
predicted fracture pressure.

6
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

History and Development of Oilfield Fracture Pressure Theory and Application

The development of fracture pressure theory and application is marked by considerable

confusion and misunderstanding, particularly in relation to the specific type of fracture
pressure and deformational behavior. For example, some authors’ objectives were to
discuss hydraulic fracturing of sands, but their conclusions and derived parameters were
inappropriately used by others to estimate fracture initiation pressures in shales. Some
authors have considered all rocks as behaving in an elastic manner, and therefore
characterized all rocks as possessing a Poisson’s ratio. Many of the authors associated
with this subject published equations and parameter values still in use today; some of
these authors’ works are relevant and appropriate today, others are not.

Twelve papers are reviewed in the following section. The starting point for any
discussion on this subject is Hubbert and Willis, whose classic 1957 paper focused on
fracture injection pressure; they detailed fracture pressure theory from the perspective of
rock mechanics. In 1967, Matthews and Kelly defined matrix stress coefficient as
representing a varying proportion of the vertical effective stress, and demonstrated that
their stress ratio was a function of depth and area. Matthews and Kelly’s paper was one
of the earliest explanations of the equivalent depth method, incorporated to reflect the
authors’ belief that the stress ratio of an abnormally-pressured formation was equivalent
to that of a normally-pressured formation. Eaton emphasized the importance of viewing
both pore and fracture pressure as a function of variable overburden gradient, and
disagreed with the equivalent depth concept as applied to fracture pressure theory.
Although Eaton confused the ideas of Poisson’s ratio and stress ratio, his 1968-dated
graphs of Poisson’s ratio versus depth and overburden gradient are still in use today. In
1972, Christman discussed the influence of water depth on fracture gradient, and
attempted to explain the differences between predicted and measured gradients as a
function of formation density. Anderson, Ingram and Zanier, in a 1973 paper
concerning the fracture initiation pressure of Gulf Coast sands, related the variation in
fracture gradients to shale content; they described a means to determine shale content
from the sonic and density log, and to derive Poisson’s ratio from this value. Pilkington
(1978) summarized the work of many authors, producing a graph and equations of stress
ratio as a function of overburden gradient, and also attempted to explain the differences
between measured and predicted fracture gradients. In one of the more enlightened
papers on this subject, Cesaroni, Giacca and Schenato described, in 1981, AGIP’s
approach, which recognized both elastic and plastic deformational behavior in rocks.
Daines (1982), however, viewed all rocks as behaving in an elastic manner, produced a
comprehensive table of Poisson’s ratios, and emphasized the necessity of recognizing
superposed tectonic stress from leak-off tests. Also in 1982, Breckels and van Eekelen,
using a worldwide database of hydraulic fracturing and leak-off values, derived equations
for determining the minimum horizontal stress. Marsh and Smith (1984) related the
methods used by Shell and Amoco; Simmons and Rau (1988) cited Chevron’s
deepwater application of a modification of Eaton’s fracture prediction method. Finally,

7
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

History and Development of Fracture Pressure Theory and Application (cont’d)

Eaton and Eaton (1997) published updated graphs and equations for Poisson’s ratio as a
function of overburden gradient for the deepwater Gulf of Mexico.

As the following papers are reviewed, it is important to bear in mind certain fundamental
points of discussion. These are:

1. Fracture injection pressure is lower than fracture initiation pressure; it may

range from less than ten to greater than fifty per cent lower.
2. If pre-existing fractures are penetrated by a borehole, the highest mud density
that can be used in that open hole cannot exceed the fracture injection
pressure.
3. Shales tend to respond plastically to an applied stress; sands tend to deform
elastically to an applied stress. Poisson’s ratio is a property of elastic rocks
only.
4. Casing shoes are set in clay/shales. In the majority of these cases, the fracture
initiation gradient of the clay/shale is equivalent to the overburden gradient.
5. If sands are exposed in the open borehole, they may tend to fracture at a
pressure generally ranging from thirty to fifty per cent lower (plus the pore
pressure) than a clay/shale.
6. For well planning purposes, fracture initiation pressure is used to plan
maximum mud weights and casing setting depths. Contingency plans are
normally incorporated in the event that pre-existing fractures are penetrated.

8
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Hubbert and Willis

Hubbert, M.K., and Willis, D.G., “Mechanics of Hydraulic Fracturing,” Journal of

Petroleum Technology, July 1957, pp. 153-163.

Hubbert and Willis (Shell) specifically set about to examine pressures related to hydraulic
fracturing. Using both soil and rock mechanics theory, they further examined the
superposed stresses caused by the presence of a borehole and reduction in stress caused
by the presence of pre-existing fractures. Although recognizing that rupture pressure
included the rock’s tensile strength and borehole stress field, they argued that in most
cases the borehole penetrates pre-existing earth fractures. These fractures, by theory,
were near vertical and perpendicular to the least horizontal stress. Consequently, the
pressure most relevant to fracturing was the injection pressure. The authors concluded
that the pressure to extend a pre-existing fracture was slightly greater than the least
horizontal stress σ3.

The authors offered evidence from rock mechanics indicating that normal faults
frequently occurred when the least horizontal compressive stress σ3 was one-third the
vertical effective stress σ1. Since the pressure to inject fluid into a pre-existing fracture
was only slightly greater than the least horizontal stress, then the fracture injection
pressure was equivalent to:

F = (S-p)/3 + p

They reduced this expression to:

F = (S + 2p)/3

For Gulf Coast normal hydrostatic conditions (i.e., normal pore pressure gradient of
0.465 psi/ft) and assuming a constant overburden gradient of 1.0 psi/ft, Hubbert and
Willis showed that the minimum fracture gradient was 0.64 psi/ft.

Hubbert and Willis considered that S, the normal (acting perpendicular to a horizontal
plane) total stress is essentially equal to the weight of the overlying sediments, which is
more commonly termed the overburden pressure. σ is the vertical component of the solid
stress, usually referred to as the effective stress. P is the pore pressure (Terzaghi’s
neutral stress); p cannot produce a shearing stress and consequently cannot cause
deformation.

9
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Hubbert and Willis (cont’d)

Discussion:

The primary intention of Hubbert and Willis was to determine an expression for a
minimum, or injection, fracture pressure, based upon sound theory and empirical results
of laboratory tests. They achieved this, to some degree, in this now-classic publication.
However, their conclusion that a fracture will occur when a pressure exceeding one-third
to one-half the vertical effective stress was subsequently criticized by other researchers.
By assuming this constant condition, they essentially said that fracture (injection)
gradient is constant with depth, increasing only with pore pressure. Subsequent
researchers observed that this hypothesis was not consistent with field data. However,
when it is borne in mind that Hubbert and Willis were concentrating on sands, their
espousal of a what was effectively a constant stress ratio (equal to 0.33) was not that
unreasonable, given an approximate Poisson’s ratio in these sands of 0.25. Although
later researchers such as Matthews and Kelly, Eaton, Pilkington, and others showed the
stress ratio (and/or Poisson’s ratio) of Gulf Coast sands was highly variable, Hubbert and
Willis did note in their paper that the least horizontal stress ranged from one third to one
half the vertical effective stress.

Hubbert and Willis’ fracture theory deals only with fracture injection pressure. In the
case where pre-existing fractures are present, mud density cannot be raised higher than
the equivalent injection pressure. However, where pre-existing fractures do not exist, the
highest mud density may be considerably greater than the fracture injection pressure.

A limitation of their fracture pressure relationship is the assumption of a constant

overburden gradient equal to 1.0 psi/ft, an assumption that was later shown to be
incorrect. Interestingly, in their reply to Reynold’s discussion of their 1957 paper,
Hubbert and Willis stated that overburden gradient was a function of sediment density
and depth.

10
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Matthews and Kelly

Matthews, W.R., and Kelly, J., “How to Predict Formation Pressure and Fracture
Gradient,” Oil and Gas Journal, Feb. 20, 1967, pp. 92-106.

In 1967, Matthews and Kelly (Mobil) proposed a fracture gradient equation relating
fracture pressure to pore pressure and some portion of the matrix pressure. They defined
this portion as the matrix stress coefficient ki, stating that ki related the actual matrix
stress to the conditions of matrix stress that would exist if the formation were normally
pressured.

They presented an equation for fracture pressure expressing that a fracture was induced
when an applied pressure equaled the matrix stress coefficient times the effective vertical
stress, plus the pore pressure:

F = k iσ + p

In their paper, the authors presented a plot of variable matrix stress coefficient versus
depth for normally-pressured formations in the South Texas Gulf Coast and the Louisiana
Gulf Coast. These values were derived by back-calculation from empirical fracture
initiation tests. This plot showed values of matrix stress coefficient ranging from
approximately 0.33 to 1.0. In the absence of local data, they said, these matrix stress
coefficient curves could be used to determine fracture pressure at any depth. The authors
made a point, however, that actual fracture data for an area are needed before a depth
versus ki curve can be established.

Matthews and Kelly stated that to determine fracture pressure in an abnormally-pressured

formation, the matrix stress coefficient must be derived from a “depth of interest” that is
normally pressured. They gave an example of this procedure, which later came to be
called the “equivalent depth” or “depth of seal” method.

The authors also made a point of recognizing that in areas of pre-existing fractures, the
pressure to open or extend these fractures is less than the fracture initiation pressure.

Discussion:

Using field data from fracture pressure tests, Matthews and Kelly demonstrated the
variability and range of the matrix stress coefficient (which we can assume they equated
with the ratio between vertical and horizontal effective stresses) with depth and area.

11
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Matthews and Kelly (cont’d)

Given their matrix stress coefficient and their fracture pressure equation, Matthews and
Kelly demonstrated that fracture gradient varied with depth and area.

For determining the matrix stress coefficient for an overpressured formation, Matthews
and Kelly stated that the matrix stress in question had to be adjusted to a “depth of
interest”. What they are advocating is a method that was later to called the “equivalent
depth”, or “depth of seal”, method. This method used some terms that differed from
those used by these authors, and to avoid any confusion, the terms of the equivalent depth
method are used in this discussion. The equivalent depth method states that the matrix
pressure (in pressure terms, not gradient or equivalent mud weight) of an abnormally-
pressured formation at depth y (the depth of interest) is equal to the matrix stress pressure
of a normally-pressured formation at depth x (the equivalent depth). Therefore, by
plotting a parameter such as matrix pressure versus depth and establishing a normal
compaction trend, an equivalent depth can be found for an abnormally pressured
formation. Then, given that the matrix stress pressure of the abnormally-pressured
formation is the same as that of the normally-pressured formation, it follows that the
matrix stress coefficient of both formations is the same. The matrix stress coefficient at
the equivalent depth is read directly from Matthews and Kelly’s plot or from a similarly
derived plot.

Although their use of a constant overburden gradient of 1.0 psi/ft was at that time
acceptable, it no longer is, and overburden gradient must be considered a variable, not a
constant. Unfortunately, what Matthews and Kelly were really dealing with were two
variables (overburden and matrix stress coefficient), each changing with depth, while
they were considering only one.

Matthews and Kelly provided correlations of ki versus depth for Texas and the Louisiana
Gulf coasts. These two curves, while showing similar trends, have different values. The
authors made no attempt to explain difference (it is due to the higher clay content of
sands offshore Texas). The authors also didn’t discuss any details of the empirical data
used to construct their correlation; they stated that the information came from fracture
initiation tests, but they did not identify whether these were leak-off tests or hydraulic
fracturing procedures. Hence, doubts exist as to the correctness of their conclusions. For
example, did they derive their matrix stress coefficient from the leak-off pressure, the
rupture pressure, or the injection pressure? Were they consistent in the selection, or were
the pressures intermingled? Were the pressure tests performed only in sand? How many
people have used these charts over the years believing they applied to shale?

12
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Eaton

Eaton, B.A., “Fracture Gradient Prediction Techniques and their Application in Drilling,
Stimulation, and Secondary Recovery Operations,” 43rd Annual SPE Meeting, Houston,
TX, Sept. 1968.

Eaton (Conoco) reviewed methods used to calculate hydraulic fracture gradients, and
presented his own technique which he believed would be more accurate and reliable.
This new technique, based on Hubbert and Willis, employed a variable stress ratio.
Eaton stated that he would show that overburden pressure, pore pressure and Poisson’s
ratio (apparently the stress ratio he referred to above) were the variables that control
fracture pressure.

Eaton began his review by stating that he sought to formulate an equation that would
predict fracture initiation and/or extension pressure. Noting that it was impossible to
know if a wellbore penetrated pre-existing fractures, he said it was similarly impossible
to differentiate between these data (referring to fracture pressure test data from which
correlations, such as stress ratio, are derived).

Eaton stated that Hubbert and Willis’ (1957) fracture extension gradient formulation
relating overburden, pore pressure and stress ratio, given by

F = (S-p)(µ/1-µ) + p

was incorrect, because they assumed the overburden gradient S was constant (1.0 psi/ft)
and the stress ratio µ (which Eaton also defines as Poisson’s ratio) was also constant
(0.25). Eaton argued that Hubbert and Willis’ formulation was incorrect because of the
evidence shown by field data, citing that fracture pressure gradient typically increased
with depth, and that actual results were commonly much higher than computed results.

In his review of the other most widely used formulation, that of Matthews and Kelly
(1967), which is:

F = k iσ + p

where ki is their variable matrix stress coefficient, Eaton said this formulation contained
two weaknesses, one being constant overburden, the other being that the stress ratio of an
abnormally-pressured formation was the same as that of the deepest normally-pressured
formation.

13
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Eaton (cont’d)

Eaton’s Technique

Eaton postulated that the Hubbert and Willis formulation was valid, but that all the
variables are functions of depth, giving his formulation as:

F = (S-p)(µ/1-µ) + p

where:

S = overburden gradient
p = pore pressure
µ = stress ratio, or Poisson’s ratio

and where overburden gradient S and stress ratio (or Poisson’s ratio) µ are functions of
depth.

Using density logs from Gulf Coast wells, Eaton derived an average variable overburden
gradient versus depth curve. Using field data from fracture tests, he also derived a
variable Poisson’s ratio curve versus depth for the Gulf Coast. By way of examples and
comparisons, Eaton demonstrated that his new technique produced more accurate fracture
gradient estimates. He stated this technique was applicable to other areas, cautioning that
local overburden gradient and Poisson’s ratio curves had to be determined.

Discussion:

Eaton concluded that overburden gradient and stress ratio were functions of depth (which
by consequence makes stress ratio a function of overburden gradient), and by doing so
greatly improved the ability to make more accurate estimations of subsurface pressures.

In his criticism of previous fracture gradient prediction techniques, he noted Hubbert and
Willis had focused on fracture extension gradient, and Matthews and Kelly on fracture
breakdown (initiation) gradient, which is greater than the fracture extension (also termed
propagation) gradient. Eaton says it is impossible to know if pre-existing fractures are
present in a formation, so it is not possible, from field data (apparently during a leak off
or fracture test, which are two very different types of tests) to distinguish the difference
between extension and initiation. Eaton states his technique will produce a fracture
pressure equivalent to fracture initiation and/or extension. So he is trying to have it both
ways, while subsequently advising that the fracture extension gradient should never be

14
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Eaton (cont’d)

exceeded while drilling. If all this sounds somewhat confusing, it is because it is a

characteristic of the author (Eaton).

Eaton credits Hubbert and Willis with the first formulation of a fracture gradient equation
incorporating, in Eaton’s words, a “stress ratio”, which Eaton identifies with the symbol
µ, and which he alternately refers to as either stress ratio or Poisson’s ratio. If one refers
to Hubbert and Willis’ original paper, they never make mention of stress ratio or
Poisson’s ratio. What prompted Eaton to put these words in their mouths can only be
guessed, but it certainly did lend credibility to Eaton’s use of Poisson’s ratio. Eaton
clearly does not understand what Poisson’s ratio is. He consistently confuses it with
Matthews and Kelly’s stress ratio. This pattern became typical of this whole subject in
later years, with some authors favoring the use of Poisson’s ratio, some the stress ratio,
others both, and some with other ideas entirely.

Eaton stated that he derived his Gulf Coast variable overburden and variable Poisson’s
ratio curves from field data, but provided no details concerning that data. This was
unfortunate, because subsequent users of his formulation very likely used it in error. It is
likely Eaton’s fracture data were related to sands, not shales. Hence, his plot of Poisson’s
ratio is a function of depth skewed for sand. Finally, because he provided no information
related to his field data, it is impossible to know whether the pressures were leak-off,
rupture, extension, or a mixture of these.

15
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Christman

Christman, S.A., “Offshore Fracture Gradients and Casing Setting Depths,” 47th Annual
SPE Meeting, San Antonio, TX, Oct 1972.

Christman (Humble) discussed the influence of water depth on fracture gradient. This
effect, he demonstrated, was to lower its magnitude, because water depth lowered one of
the primary controlling variables, overburden gradient. Christman aimed to develop a
fracture gradient technique for offshore use.

Christman reviewed previous literature and determined that the most important
parameters affecting fracture initiation pressure were overburden gradient, pore pressure,
and the horizontal-to-vertical stress ratio, while parameters such as rock compressibility
and Poisson’s ratio had relatively little influence. Of the influential parameters, he felt
the stress ratio k was the most significant, yet elusive, variable. However, he felt that
Matthews and Kelly’s hypothesis that stress ratio of an abnormally-pressured formation
was the same a shallower, normally-pressured formation of the same matrix pressure was
reasonable.

Christman’s fracture gradient formula was:

F = kσ + p

where:

k = horizontal-to-vertical stress ratio

His technique involved determination of variable overburden gradient S from offset well
density logs, pore pressure p from the best applicable techniques, and stress ratio k from
well fracture tests.

Christman noted that his technique sometimes predicted fracture gradients that were
lower, sometimes higher, than actual. He also noted that where predictions were higher
than actual coincided with low density formations, and where predictions were lower
than actual with high density formations. He postulated an alternate stress ratio
correlation based on formation density. Using density logs and well data, he plotted
stress ratio versus density; the relationship was linear, with stress ratio increasing with
increasing density. Christman felt this technique provided for more accurate fracture
pressure estimations and should be applicable elsewhere.

16
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Christman (cont’d)

Christman showed the effect of water depth on offshore overburden gradients. As water
depth increased, fracture gradients were considerably reduced. He noted this effect was
particularly significant at shallow depths, and consequently had a major influence on
casing setting depths.

Christman also noted that an elevated mud return line caused a reduction in fracture
gradient, through its influence on mud hydrostatic pressure above subsea (mean sea level)
datum. He stated that this “air gap” effect could be significant, particularly at shallow
penetrations. He provided an example of a flowline elevated 75 feet above a mean sea
level datum. A fracture pressure of 520 psi at 1000 feet subsea gave an equivalent
fracture mud weight of 10.0 ppg. But the 75 foot air gap meant a hydrostatic head of
1075 feet, such that the formation at 1000 feet subsea will fracture with a 9.3 ppg mud.

Discussion:

Christman, like Matthews and Kelly and Eaton before him, recognized the influence of a
variable stress ratio on fracture gradient estimations. Christman, like Eaton, also
recognized the significance of the variable character of overburden gradient. His
technique, similar to Matthews and Kelly’s and Eaton’s, involved determination of local
stress ratios from fracture pressure tests. Like those authors, Christman did not specify
either the type or source of his data. Christman recognized that efforts to determine
correlations of stress ratio versus depth sometimes did not produce accurate fracture
pressure predictions, and he surmised this may be due to formation density. Using
density log and fracture test data, he showed an excellent correlation between stress ratio
and bulk density.

Christman may have been on to a more improved corrrelation for stress ratio, i.e., stress
ratio versus bulk density, as opposed to stress ratio versus depth - but he did not offer a
means of using these results for fracture pressure predictions. And this is the problem: it
would be necessary to know the formation bulk densities exposed in the wellbore.

Christman demonstrated the effect of water depth on offshore fracture gradients. This
effect may indeed present significant problems, particularly at shallow depths. For
example, in deepwater projects, shallow penetration depths are drilled with returns to the
seabed because attempts to return cuttings to the surface would both fracture and wash
the wellbore.

Christman discusses the “air gap” effect of offshore flowlines. This effect is to reduce
the equivalent mud weight of fracture gradients, and this reduction may be very large,
particularly at shallow depths. This effect, however, does not exist if the drill floor, as

17
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Christman (cont’d)

opposed to mean sea level, is the reference datum. Also, it is important not to confuse
this air gap with the air gap that applies to overburden gradients.

18
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Anderson, Ingram and Zanier

Anderson, R.A., Ingram, D.S., and Zanier, A.M., “Fracture Pressure Gradients from Well
Logs,” Journal of Petroleum Technology, Nov. 1973, pp. 1259-1268.

Anderson, Ingram (Schlumberger) and Zanier (Conoco) developed a theoretical basis to

support their contention that mechanical properties have a strong influence on fracture
pressures. They noted that fracture pressures often varied at the same depth within the
same field, that some sands were weak, other strong. As further evidence, they cited
Matthews and Kelly’s stress ratio versus depth relationship for the Gulf Coast, showing a
distinct difference between South Texas and Louisiana. The authors pointed out that
techniques such as Hubbert and Willis, Matthews and Kelly, and Eaton either ignored
formation properties or assumed them to be a relatively monotonic function of depth.

The authors identified Poisson’s ratio as one of the most important parameters accounting
for variations in formation properties. They recognized and compared fracture gradient
formulations based upon Terzaghi and Biot:

F = (2ν/1-ν)S + (1-3ν/1-ν)p Terzaghi

F = (2ν/1-ν)S + α(1-3ν/1-ν)p Biot

where:

ν = Poisson’s ratio, = λ/2(λ + µ), where λ and µ are the Lame constants
α = 1- Cr/Cb, where: Cr = solid rock compressibility,
Cb = porous rock compressiblity

and both Terzaghi’s and Biot’s stress strain relationships are applicable for elastic,
isotropic, porous media.

The authors noted the similarity in these two equations and chose to use Biot’s because
the pore pressure term disappears at zero porosity. They stated, though, that they could
have just as easily pursued this same question by examination of the ratio of the
horizontal to vertical stress, such that the above questions become:

F = 2kS + (1-2k)p

F = 2kS + α (1-2k)p

19
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Anderson, Ingram and Zanier (cont’d)

In this latter case, matrix stress coefficient would be correlated with shale content. The
authors note the problem then is to determine a relationship between Poisson’s ratio and
shale content. Shale content can be determined from sonic and density logs:

Ish = φS - φD/φS

where:

Ish = shale content

φS = porosity from sonic log
φD = porosity from density log

The authors mentioned that other shale indices could also be used. To calculate Poisson’s
ratio, they rearranged Biot’s equation:

ν = (F - αp)/(F + 2S - 3αp)

However, it was first necessary to evaluate the α parameter. Given its relationship to
porosity, they approximated it by:

α = 1 - (1- φD )

which simplifies to α = φD . Then, using 29 sets of fracture data from the Gulf Coast,
they plotted Poisson’s ratio versus shale content, which showed a linear relationship.

The authors concluded that they had developed an important empirical relationship
between Poisson’s ratio and shale content of Gulf Coast sands, enabling improved
fracture pressure estimates. Shale content was relatively easily derived from sonic and
density logs, which also provided means for determining pore and overburden pressures.

Discussion:

The authors attempted to explain the differences observed in fracture pressure tests
performed at the same depths within the same field, and also differences observed in
matrix stress coefficients in different parts of the Gulf of Mexico. They reasoned the
shale content of sands influenced Poisson’s ratio of the sands, with the shale “acting as a
plastic bonding agent.” Using 29 sets of fracture pressure data from twelve Gulf Coast
wells, and attempting to use initiation pressure unless propagation pressure “was not

20
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Anderson, Ingram and Zanier (cont’d)

significantly lower,” they back-calculated Poisson’s ratio, which they then correlated
with shale content in the sands. The resulting correlation was linear, showing Poisson’s
ratio ranging from approximately 0.27 at 0% shale content to approximately 0.37 at 50%
shale content. This correlation, as the authors were quick to point out, was valid only for
Gulf Coast sands.

Like others before them, Anderson et al. focused on sands, not shales, and like Matthews
and Kelly, Eaton, etc., they identified that Poisson’s ratio or stress ratio was a significant
controlling determinator of fracture pressure. Other authors had conjectured that
shaliness explained the observed differences in fracture pressures. Christman, for
instance, derived a linear correlation between stress ratio and bulk density.

A limitation to Anderson et. al.’s technique is that it applies only to sands, and cannot be
applied for determining the fracture initiation pressure of shales.

21
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Pilkington

Pilkington, P.E., “Fracture Gradient Estimates in Tertiary Basins,” Petroleum Engineer

International, May 1978, pp. 138-148.

Pilkington (Conoco) developed a predictive fracture gradient equation for rank wildcat
use in Tertiary basins, based on variable overburden gradient and stress ratio. He
reviewed the equations and techniques developed by Hubbert and Willis, Matthews and
Kelly, Eaton, Pennebaker, and Christman. He specifically noted that it was Pennebaker
who believed overburden gradient was the controlling factor for stress ratio in relaxed
Tertiary basins. Using these authors’ data and correlations, Pilkington constructed a
graph of average stress ratio ka versus depth, a graph and the following equations relating
ka to overburden gradient:

ka = 3.9 (S/D) - 2.88 for ((S/D) ≤ 0.94)

ka = 3.2 (S/D) - 2.224 for ((S/D) > 0.94)

Pilkington constructed a plot of fracture gradients versus depth using his average stress
ratio values and published data. This plot showed considerable scatter; he attributed this
scatter to a number of influences, including:

1. Some fracture gradients were derived from propagation, not initiation,

pressures.
2. Data from some leak-off tests were possibly affected by poor cement jobs.
3. Some values were not corrected for water depth.
4. Some results were influenced by mud weights (muds with high mud weights
may exhibit low penetrability, those with low weights often penetrate into the
formation, resulting in lower fracture pressures).

Pilkington described the use of his technique in wildcat areas:

1. Calculation of variable overburden gradient from the density log.

2. Use the graph of average stress ratio versus overburden gradient to determine
the average stress ratio ka; if a local data are available, construct a similar
correlation.
3. With a good estimate of pore pressure, calculate fracture pressure gradient
using F = ka σ + p.

22
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Pilkington (cont’d)

Pilkington provided examples and concluded that the accuracy was +/- 0.5 ppg. Finally,
he reiterated that this and similar techniques did not apply to evaporites, brittle shales,
older rocks, and tectonically stressed areas. He also noted that at shallow depths, fracture
gradient often was approximately equal to overburden gradient because the least principal
stress was vertical (i.e, overburden gradient was low).

Discussion:

Although Pilkington was convincing in synthesizing the work of previous authors, his
results, in the end, were not appreciably better. He attempted to improve results by
correlating stress ratio with overburden gradient, which in essence is the very similar to
Christman’s correlation of stress ratio with bulk density. He gave as his reasons for
scatter in results using his method with other authors data as being due to factors such as
propagation versus initiation pressures, values that were not corrected for water depth,
etc. Interestingly, he did not attempt to invoke any other factors that may influence the
stress ratio. For example, Pilkington shows Matthews and Kelly’s stress ratio versus
depth plot for Texas and Louisiana, but does not attempt to explain the differences. One
would be left to assume, given his explanations, that the differences must be due to
propagation or initiation pressures, water depth, or drilling fluid characteristics. It is
obvious that one area’s stress ratios would not be skewed this much from another area’s
solely on the basis of these factors. So Pilkington failed to recognize the influence of
shaliness (in sands) on fracture pressure.

Examination of Pilkington’s stress ratio equation ka = 3.9 (S/D) - 2.88 reveals that it will
return a zero when the overburden gradient is equivalent to 14.2 ppg. This means that it
cannot be used in many offshore situations.

23
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Cesaroni, Giacca, Schenato, and Thieree

Cesaroni, R., Giacca, D., Schenato, A., and Thieree, B., “Determining Fracture Gradients
While Drilling,” Petroleum Engineer International, June 1981, pp. 60-86.

Cesaroni, Giacca, Schenato, (AGIP) and Thieree (GeoServices) presented a

comprehensive examination of the techniques used by AGIP to estimate overburden, pore
and fracture pressures. They presented equations for fracture pressure estimation that are
a function of mechanical behavior, lithology, and drilling fluid penetrability.

The authors distinguished three different fracture pressure equations:

1. For elastic behavior, applicable to lithologies such as sands and sandstones, and for
non-penetrating drilling fluids (i.e., where differential pressure between mud and
formation is totally withstood by either the mud cake or low formation permeability):

F = (2ν/1-ν)(S - p) + p

2. For elastic behavior and penetrating drilling fluids:

F = 2ν(S - p) + p

3. For plastic behavior, lithologies such as shales, marls, salt, etc.:

F=S
where:
ν = Poisson’s ratio
and:
ν = 0.25 for clean sands, sandstones, unfractured carbonates, at shallow
depths
ν = 0.28 for shaley sands, sandstones and carbonates at greater depths

The authors provided no underlying theory or explanations for their fracture pressure
equations.

Discussion:

The authors have extensive experience in deep, hot, onshore and offshore drilling in
clastic and carbonate environments. They discussed pore pressure derived from the

24
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Cesaroni, Giacca, Schenato, and Thieree (cont’d)

Sigmalog, and variable overburden gradients estimated from sonic and density logs.
They correctly recognized Poisson’s ratio as a mechanical property of elastic rocks. They
believed the range of Poisson’s ratio was essentially limited, from 0.25 to 0.28. They felt
that Poisson’s ratio was not applicable to rocks that exhibited plastic behavior. They did
not, however, attempt to establish criteria to distinguish elastic from visco-elastic or
plastic behavior.

There exists some precedence for their beliefs. Heim, and some more recent researchers,
provided evidence or convincing arguments for the extension of the range of Poisson’s
ratio from 0.30 to 0.50. It has often been observed in shales that the fracture pressure is
approximately equal to overburden pressure; if the behavior of these shales was
considered elastic and their Poisson’s ratio equal to 0.50, then the equation F = (ν/1-ν)(S
- p) + p simplifies to F = S.

Of the various techniques discussed, the AGIP formulations are perhaps the most
enlightened. Certainly for most deepwater clay/shales, the fracture initiation gradient is
equivalent to the overburden gradient.

25
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Daines

Daines, S.R., “Prediction of Fracture Pressures for Wildcat Wells,” Journal of Petroleum
Technology, April 1982, pp. 863-872.

Daines (EXLOG) stated that Gulf Coast empirical formulas for fracture pressure were
unable to make accurate predictions outside that area. He believed two factors were
responsible for this inaccuracy: empirically-derived stress ratios which are inapplicable to
areas outside the Gulf, and the failure to recognize superposed tectonic stress.

To address the first problem, Daines identified Poisson’s ratio as a primary mechanical
property of rocks that influences their response to stress. He noted that Poisson’s ratio
was unique for each and every lithology, and he provided a table of values
experimentally measured using sonic testing. This table included a spectrum of
lithologies, with Poisson’s ratio ranging from as low as 0.01 for fossiliferous sandstone to
0.50 for wet clay.

In terms of subsurface state of stress, Daines argued from the premise that the horizontal
principal stress is a function of effective vertical stress and Poisson’s ratio, viz:

σH = σ (ν/1-ν)

where:

σH = horizontal effective stress

σ = vertical effective stress
ν = Poisson’s ratio

Daines argued from Hafner’s (1951) perspective, that any area may be subject to some
tectonic stress that would be superposed and constant in any horizontal plane and
increases uniformly with depth. Daines said this superposed stress may range in
magnitude from zero to three times the minimum horizontal stress. Consequently, the
minimum horizontal stress was expressed as:

σH = σt + σ (ν/1-ν)

where σt is the superposed horizontal stress, and σt / σ = B, which is constant with

depth.
The equation for fracture pressure is then:

F = σt + σ (ν/1-ν) + p

26
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Daines ( cont’d)

Daines stated that the superposed tectonic stress could be determined from the first
fracture test in a compact formation:

σt = F - [σ (ν/1-ν)] + p

Daines noted that many factors affected fracture tests, including:

1. Higher mud weights and mud gels, smaller hole diameters, and higher pump rates
produced higher fracture pressures.
2. Hole deviation significantly affected fracture pressures.
3. Sensor instrumentation limited fracture pressure accuracy to 5%.
4. Mud penetrability did not alter breakdown pressures but may affect the pressure plot
to the point where the key point is obscured.

Daines concluded that an equation and technique had been developed to determine
stresses - both tectonic and minimum horizontal - which when used with accurate
estimates of pore pressure and overburden pressure and Poisson’s ratio for specific
lithologies, provided a more accurate fracture pressure prediction in wildcat areas than
the Gulf Coast empirically-derived formulas. These fracture pressures were predicted
with an accuracy of 95%.

Discussion:

Daines, like Anderson et. al, believed that in many geographic areas additional tectonic
stresses are superposed upon the gravity-influenced horizontal principal stresses. Daines,
however, believed that he could, in a practical manner, separate and identify the tectonic
stress σt from total stress, assuming the ratio of the tectonic to the minimum horizontal
stress remained constant with depth. Combined with a unique Poisson’s ratio for each
lithology, Daines firmly believed his equation and technique would produce more
accurate fracture pressure estimates than the empirically-derived and basin-specific
parameters of the Gulf Coast.

Daines uses Poisson’s ratio values determined using sonic testing, and these values range
from 0.01 to 0.50 for lithologies including sediments, metamorphics and igneous rocks.
For sands alone, the values range from 0.01 to 0.24; for clays, from 0.17 to 0.50. Daines
presumes that all lithologies behave in an elastic manner.

Daines assumptions and conclusions are open to argument and dispute. Many
researchers

27
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Daines (cont’d)

have stated that sonic-derived values of Poisson’s ratio are subject to much error. Other
researchers do not accept that all rocks will behave in an elastic manner. Daines argued
that Poisson’s ratio and stress values derived from the Gulf Coast introduced error when
used in other areas, and if this argument is accepted, the question has to be asked where
he sourced his values. It is highly doubtful that values are available for wildcat areas, so
on a practical basis, error must be introduced when Daines’ values are used for predictive
fracture pressure estimates.

Daines maintains that tectonic stress must be identified at a first fracture test and
thereafter used in all subsequent fracture pressure predictions. However, if a user selects
any other than the exact Poisson’s ratio for the lithology in a fracture test, a tectonic
stress may be identified that does not actually exist. All subsequent predictions, then,
will be subject to this error.

Given the manner in which Daines recognizes an extremely large range in Poisson’s
ratio, and his insistence that all lithologies deform in an elastic manner, it would be
expected that his technique would usually yield a value for superposed tectonic stress,
even in its absence. This is because most fracture tests performed during drilling are
leak-off tests, and these tests will tend to measure the borehole hoop stresses. It is not
inconceivable that these stresses would be misidentified as a superposed tectonic stress.

28
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Breckels and van Eekelen

Breckels, I.M. and van Eekelen, H.A.M., “Relationship between Horizontal Stress and
Depth in a Sedimentary Basin,” Journal of Petroleum Technology, Sept. 1982, pp. 2191-
2199.

Breckels and van Eekelen (Shell) developed a correlation of minimum horizontal stress
versus depth and pore pressure for the Gulf Coast and other Tertiary basins. They refined
their Gulf Coast correlation for general worldwide use.

Breckels and van Eekelen distinguished between fracture initiation and fracture
propagation pressures. They noted that most earlier researchers ignored this difference,
primarily because their fracture estimation formulations focused on the use of a stress
ratio representing the minimum horizontal stress divided by the effective (vertical) stress.
They pointed out that stress ratio, when derived from leak-off tests, needed to be
multiplied by a factor ranging in value from 1 to 2, to account for the difference between
the two horizontal stresses and a potential hoop stress effect.

The authors stated that the instantaneous shut-in pressure from a hydraulic fracturing job
is the closest approximation to the minimal horizontal stress. But because most fracture
tests are performed for well design and planning purposes, they explained a means to
approximate the minimum horizontal stress from leak-off tests.

The authors state that the minimum horizontal stress cannot be derived from fracture
breakdown tests, because breakdown pressure pb is:

pb = 3SHMIN - SHMAX - p

where SHMIN is the minimum horizontal stress, SHMAX is the maximum horizontal stress
and p is the pore pressure. Furthermore, breakdown pressure is influenced by hole
geometry, hole integrity, mudcake properties, and the extent to which fluid penetrates
into the formation. Therefore, the only way to estimate the minimum horizontal stress is
to use leak-off test data. However, the fracture initiation pressure may be affected by
hoop stress, such that:

SHMIN ≤ pf ≤ 2SHMIN - p

where pf is the fracture initiation pressure.

29
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Breckels and van Eekelen (cont’d)

Using data only from hydraulic fracturing tests, the authors derived plots of SHMIN versus
depth for the Gulf Coast, Venezuela, Brunei, the North Sea, and The Netherlands. Their
analysis showed:

1. Some differences between these plots for different basins can be minimized
when tectonic stress differences are taken into account.
2. There existed, in a few basins, differences in SHMIN versus depth for different
rock types; these differences ranged from statistically insignificant to
potentially significant.

The authors observed that the difference between SHMIN derived from leak-off tests and
SHMIN derived from instantaneous shut-in pressures from hydraulic fracturing tests
averaged approximately 11%. Making this adjustment, they derived equations for SHMIN
that can be used in tectonically relaxed areas such as the Gulf of Mexico, North Sea, and
The Netherlands:

SHMIN = 0.197D1.145 + 0.46(P-Pn) for D < 11,500 feet

SHMIN = 1.167D - 4596 + 0.46(P-Pn) for D > 11,500 feet

Discussion:

The authors, using fracture pressure (leak-off) and hydraulic fracturing data, and
distinguishing between these two, have derived, with data from the Gulf Coast,
Venezuela, Brunei, North Sea and The Netherlands, equations for the estimation of
minimum horizontal stress that can be used in tectonically relaxed Tertiary basins. These
empirically-derived equations include the influence of formation pore pressure.

Given depth, these equations estimate the magnitude of fracture initiation pressure.
However, without a relationship keyed to overburden gradient, it is difficult to imagine
that this technique will gain widespread acceptance.

30
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Marsh and Smith

Marsh, G.L., and Smith, J.R., “Exploratory Well Design for 5,000- to 7,500-Ft Water
Depths, U.S. East Coast,” 16th Annual Offshore Technology Conference, Houston, TX,
May 7-9, 1984.

Marsh (Shell) and Smith (Amoco) discussed fracture and overburden gradient prediction
techniques used by Shell and Amoco during deepwater drilling off the U.S. East Coast in
1983.

For fracture gradient prediction, the companies used the Hubbert and Willis equation,
with stress ratio k estimated using the equation:

k = 0.3376 + 0.732 x 10-4 D - 0.2892 x 10-8 D2 + 0.4499 x 10-13 D3

For overburden gradient, the companies used an equation empirically-derived by Shell

from deep water wells:

Sv = D + 0.445Dw - 5.0

where:

Sv = overburden gradient, psi/ft

D = depth below mudline, ft
Dw = water depth, ft

The authors stated that this overburden gradient equation was more accurate than using
an equation for overburden gradient developed by Shell for the Gulf Coast. They
concluded that fracture gradients in deep water were weak but sufficiently accurate
predictions could be made to develop a sound well plan.

Discussion:

This was a light and very brief treatment in a paper that concentrated on the Wilmington
Canyon exploration program of which this was the first well. The authors stated they
were using Hubbert and Willis’s equation, but this is incorrect, for Hubbert and Willis
never used a stress ratio k. Marsh and Smith said they calculated stress ratio k from an
equation cited by Pilkington, but Pilkington never cited this equation.

31
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Marsh and Smith (cont’d)

The authors mentioned that they derived overburden gradient from an equation based
upon deep water wells. They said this equation gave better results than another equation
based on Gulf Coast data. This is puzzling, because this well was the first in the
Wilmington Canyon, and the question is, where did the data come from that was used to
develop the overburden gradient equation? Evaluation of their overburden gradient
equation shows that they consider the water gradient (0.445 psi/ft), but consider the
sediment gradient at 1.0 psi/ft. This value is clearly both wrong and inappropriate,
especially for offshore deep water locations. One must wonder why 5 psi is subtracted
from the overburden pressure.

32
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Simmons and Rau

Simmons, E.L., and Rau, W.E., “Predicting Deepwater Fracture Pressures: A Proposal,”
63rd Annual SPE Technical Conference and Exhibition, Houston, TX, Oct 2-5, 1988.

Simmons and Rau (Chevron) stated that a problem of some magnitude was the accurate
prediction of fracture gradients in deep water. In this paper, they proposed the use of two
techniques:

1. The Eaton fracture gradient equation, modified to include the effect of a column of
sea water.
2. Use of the bulk density log to produce better estimates of overburden gradient.

Modified Eaton Fracture Pressure Technique

1. Determine an equivalent sediment depth for the column of sea water, using their
graph or either of the following equations:

Deq = 1.489554 x 10-5 + [1.911364/WD + 2.703024/(WD)2]-1

or
Deq = WD/2

2. Calculate effective sediment depth to be used for overburden gradient and stress ratio
determinations. The authors stated that the effective sediment depth Deff accounted
for the pre-stressing of sediments due to the hydrostatic head of water.

Deff = Deq + Dsed

3. Determine average overburden gradient at Deff using their figure or:

Sa = 0.8511934 x e[(ln Deff - 6.206593)2/84.36084]

4. Calculate composite vertical stress σvc (overburden pressure) at the depth of interest:

σvc = (0.442 x WD) + (Sa x Dsed)

33
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Simmons and Rau (cont’d)

5. Determine matrix stress ratio k at the effective sediment depth using Eaton’s graph or
the following equation:

ke = 0.05329427 x 0.999996(Deff) x Deff(0.3006479)

6. Calculate fracture pressure:

F = ke (σvc - p) + p

In their technique, the authors noted that Deff was used to determine overburden gradient
and matrix stress ratio, which in both cases yielded higher values to account for pre-
stressing of the sediments by the column of sea water. This, they maintained, would
produce more accurate fracture pressure estimates. By comparing this modified
technique with the normal Eaton technique, their results averaged 3.38% too low versus
4.68% too high. They stated that although too high, these results were conservative and
provided an extra safety margin. However, they suggested that better accuracy would be
achieved using area-dependent overburden curves.

To develop area-dependent overburden curves, the authors proposed integration of the

bulk density log. They set out a step-by-step procedure.

Finally, the authors defined the matrix stress ratio as the ratio of the horizontal to vertical
stress, but mentioned that one generally accepted formulation (Eaton) related this to
Poisson’s ratio. Arguing that this was a property of the rock and its environment and was
therefore specific to an area, they recommended deriving local empirical curves.

Discussion:

The authors emphasized that a very big problem was the inaccuracy of fracture gradient
predictions in deep water environments. They pointed out that existing techniques did
not account for the effect of a column of sea water (offshore water depth). They
proposed a solution that made use of Eaton’s 1969 Gulf Coast variable overburden
gradient curve and variable Poisson’s ratio curve. The depth that was to be used to select
the correct values from these charts was their effective depth, which was the depth of the
sediment

34
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Simmons and Rau (cont’d)

plus one half the water depth. The authors’ proof of their contention for better fracture
pressure estimates were their results that averaged 3.4% lower than actual, as compared
to 4.7% too high without using this technique.

For all their effort, the improvement in accuracy that the authors achieved may be
considered as statistically insignificant. If as they said, the 4.7% overestimation defined a
“problem of some magnitude”, does a 3.4% underestimation define a problem solved?
And although the authors said even better accuracy would be achieved using overburden
derived from local density logs, the authors didn’t so much as provide an example.

The question has to be asked of these authors, with access to 240 sets of deep water leak-
off tests, and having in hand local density logs, why would they use Eaton’s average Gulf
Coast data?

Although the authors correctly recognized that Eaton’s curves could not be used without
making an adjustment for water depth, their solution is extremely more complicated than
is necessary. It is very arguable, for instance, that the effect of a column of seawater is
the same as half that column’s height of sediment with a density of 2.0 g/cc. The authors
provide both graphs and equations for determining the various parameters, but the
authors have no comprehension of meaningful equation development or significant
figures.

35
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Eaton and Eaton

Eaton, B.A., and Eaton, T.L., “Fracture Gradient Prediction for the New Generation,”
World Oil, Oct. 1997, pp. 93-100.

In this paper, Eaton (Eaton Consultants) and Eaton (student) reviewed how to correctly
use the Eaton methods, and presented new overburden gradient and Poisson’s ratio
curves for the Gulf Coast. The authors provided example calculations of overburden
gradient using density log values. They stated that sometimes density log data did not
exist, and recommended using seismic data to determine bulk density. They quoted a
relationship between seismic interval velocity and bulk density, credited to Gardner et. al.
(1974):

ρb = 0.23V0.25

where:

ρb = bulk density, g/cc

V= interval velocity, ft/sec

The Eatons stated there is no excuse to not develop an overburden curve for any location.
Having said this, they then presented an overburden gradient curve for the deepwater
Gulf of Mexico from a 70 well Amoco data source (Barker and Wood, 1997). They said
that this curve could be used throughout the gulf, provided the adjustment was made for
water depth.

The Eatons briefly reviewed his pore pressure equations and noted that the exponents
may be changed to suit local conditions.

The authors stated that Barker and Wood (1997) assumed that fracture gradient equaled
overburden gradient in the deepwater Gulf, and that Roche and Bourgoyne (1996) had
concluded pore pressure had no effect on fracture gradient. Using the Eaton equation, the
authors demonstrated their disagreement with these other authors. Also using the Eaton
equation, they back-calculated Poisson’s ratio for the deepwater Gulf, and presented a
graph showing Eaton’s original Poisson’s curve and the new Poisson’s curve.

Gulf Coast original Poisson’s ratio curve equations:

ν = -7.5 x 10-9 (D)2 + 8.0214286 x 10-5(D) + 0.2007142857 for D < 5000’

ν = -1.77258 x 10-10(D)2 + 9.4748424 x 10-6(D) + 0.3724340861 for D > 5000’

36
DEEPWATER PRESSURE SERVICE FRACTURE PRESSURES

Eaton and Eaton (cont’d)

Deepwater Gulf Poisson’s curve equations:

ν = -6.089286 x 10-9 (D)2 + 5.7875 x 10-5(D) + 0.3124642857 for D < 5000’

ν = -1.882 x 10-10 (D)2 + 7.2947129 x 10-6(D) + 0.4260341387 for D > 5000’

The authors opined that the new curve or equations seemingly predicted good values for a
sediment with significant amounts of clay, even outside the Gulf.

Discussion:

The Eatons continued the use of Poisson’s ratio but did not address it as other than a
function that varied with depth. Eaton never provided any details of his original
Poisson’s ratio curve, and he repeats the same here.

The Eatons suggest the use of Gardner’s equation, relating seismic velocity to bulk
density, as a means of obtaining formation density data in the absence of a density log,
for estimation of overburden pressures. However, the Eatons fail to mention that
Gardner’s equation, developed a quarter of a century ago, was derived from a limited set
of Paleocene age shale data. The equation was only a rough approximation of that data,
and there exists no credible evidence that Gardner’s equation can be applied to shales of
any other age.

37
This page intentionally left blank.
Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003

Pressure Prediction Methods

Overburden Gradient Prediction Methods

Overburden pressure is defined as a variable pressure acting vertically

downwards from the earth’s surface, the magnitude of which is dependent upon
the densities of the sediment matrices and the densities of the sediment pore
fluids. The influence of water depth and air gap must be factored into the
calculation of sediment overburden. Increasing water depth can reduce the
magnitude of overburden significantly. Several overburden gradient tables were
generated using regional overburden tables, offset wireline sonic data and a
seismic velocity profile from the prospective well to determine the best fit for the
prospective area.

Bell Regional Overburden Tables

To calculate the “initial” overburden gradient for the prospective well, Bell’s Gulf
Coast regional (Louisiana and Texas) overburden tables were used after being
corrected for the well’s air gap and water depth. While these tables provide a
good estimation of overburden, they lose some accuracy in deep water
environments.

Belloti (AGIP)

For a closer approximation, density was calculated using ∆t converted sonic

transit times (ITT) from offset well data. The following equation were used to
determine density in g/cc.

∆t (µsec/ft) = # ITT * 1000/interval thickness(ft)

ρ = 2.75 - 2.11((∆t - ∆tm)/( ∆t + ∆tf)

where: ∆tm = matrix transit time, (µsec/ft)

(sands @ 52.6 to 58.8 (50), shale @ 160 to 47 (47), salt @ 66.7)
∆tf = fluid transit time, (µsec/ft)
(sea water @ 190 to 200 (190), fresh water @ 218)

Once the average density of an interval is determined, it is converted to a

gradient by the equation below.

ρb interval (g/cc) * 0.433 => interval pressure gradient (psi/ft)

Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003
This interval pressure gradient is converted to an interval overburden pressure
and these interval overburden pressures are summed together. The cumulative
pressures are converted to an equivalent mud weight at a true vertical depth of
interest.

IPG (psi/ft) * ft (interval) => interval overburden (psi)

psi1 + psi2 + ….+ psiN => cumulative overburden (psi)

p = psi/.052 * TVD => cumulative overburden EMW (PPG)

Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003

Pore Pressure Prediction Methods

There are several factors that influence the sediment’s pore pressure. These
factors can determine the degree to which sediments can be subnormal, normal,
or abnormally pressured. This report mainly deals with recognizing abnormally
pressured sediments which have compaction-related mechanisms. Compaction-
related mechanisms include rapid deposition, high sand/shale ratio, deposition of
impermeable sediments and clay diagenesis. The main factors which influence
the magnitude of the abnormal pore pressure is the lack of hydraulic
communication, concentration of dissolved salts, presence of gas, and
magnitude of geothermal gradient. The impact on pressure of tectonic-related
mechanisms like faults and salt dome emplacement is harder to predict by
current methods. Influence of dipping beds should also be considered and
adjusted for this centroid effect (Traugott-Amoco) along with type of pore fluid
(gas, oil, or sea water) present. Other secondary abnormal pressure indicators
are also considered including connection/trip gases, drill cuttings size and shape,
torque, drag, temperature and hole fill.

This pore pressure analysis was performed using various industry recognized
petrophysical techniques using resistivity, gamma, density, sonic, and seismic
velocity data. The primary methods used are taken from Eaton’s Ratio, Alixant,
Bryant’s Dual Shale and Corrected Drilling Exponent (Dxc) calculations. Each of
these techniques is briefly explained below.
Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003
Eaton’s Ratio Method

This is probably the industry’s most recognized method to calculate pore pressure. It is based on
the assumption that the resistivity of normally compacted shales will increase with depth due to
greater compaction as depth increases. A trend line is drawn through the normally compacted or
normally pressured shales (which were assumed to be 8.6 PPG). This trend line represents what
the shale resistivities/∆t would be if they were all normally pressured. The trend line is
established by choosing a best fit straight line through the resistivity/∆t values obtained in
normally pressured shales and plotted on a semi-log graph. Shale resistivity/∆t values below the
normal trend indicate the shale is becoming under-compacted and most likely over-pressured.
The placement of this trend line is very important for accurate predictions. Eaton’s equation for
determining pore pressure based on the shale resistivities/∆t is defined as:
P = S - (S - Pn)(Ro/Rn)1.2 (resistivity)
Where: P = Pore Pressure
S = Variable Overburden For TVD
Pn = Normal Pore Pressure (8.6 PPG) SEG
Ro = Observed Resistivity in Shale
Rn = Normal Resistivity in Shale

P = S - (S - Pn)(∆tn/∆to)3.0 (sonic transit time)

Where: Pn = Normal Pore Pressure (8.6 PPG) SEG

∆to = Observed Sonic Transit Time in Shale
∆tn = Normal Sonic Transit Time in Shale

This equation can be expanded to calculate pore pressure using conductivity or

drilling exponent.

P = S - (S - Pn)(Cn/Co)1.2 (conductivity)

P = S - (S - Pn)(Dco/dcn)1.2 (Dxc exponent)

Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003
Alixant and Desbrandes

This is an explicit approach to determine pore pressure developed by

compaction-related mechanisms. The method estimates shale porosity using a
petrophysical module and effective stress in a mechanical module. This
technique must be corrected for formation temperature or regional temperature
profile. The execution of several steps must be done to derive pore pressure.
These steps are calculating bound water resistivity (Rwb), formation factor (F),
porosity (φ), void ratio ( r ), and vertical effective stress.

Rwb = 297.6T-1.76 (bound water resistivity, corrected

for BHT)

F = Rshcor/Rwb (formation factor)

F = 1 + 1.85[(1- φ)/(φ - 0.1)] (porosity)

r = φ/(1-φ) (void ratio)

σev = 10 (r-3.84)/1.1 (vertical effective stress, psi)

Once these factors are derived, the results can be inputted into the following
equation to calculate pore pressure at a measured Rsh @ TVD, formation
temperature (degrees F), and cumulative overburden (psi).

p = S - σev - 1355 = psi (pore pressure, psi)

p = psi/.052 * TVD = PPG (equivalent mud weight)

Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003

Bryant’s Dual Shale Method

This method is based on using dual shale resistivities, the observed shale
resistivity (Rsh) and a local shale water resistivity (Rw). This technique does not
require a normal trend line to be established to estimate a pore pressure like
Eaton’s equation. It is integrated utilizing Terzaghi effective stress law, Archie’s
formation factor (F) and the power law compaction relationship. The first step is
to derive shale porosity by using a re-arranged Archie’s formation factor
equation.

F = Ro/Rw = 1/φm (formation factor)

φ = exp((ln Rw - ln Rsh)/m) (dual shale equation for shale

porosity)

Where: φ = shale porosity

exp = base of Napieren logarithm
Rw = shale water resistivity
Rsh = shale resistivity
m = Archie pore tortuousity exponent
ln = natural logarithm

A Tertiary basin compaction equation then relates the actual shale effective
stress (matrix pressure) into porosity.

σact = σmax * (1- φ)7.47 (power law clay compaction

relating porosity)

Where: σact = actual effective stress

σmax = effective stress to create zero porosity in shale (@20,000 ft
= 11225 psi)
φ = shale porosity

To relate the determination of an actual shale effective stress to two shale

properties, shale pore fluid resistivity and shale bulk resistivity, the equation
below is used to derive actual effective stress ( σact) for each depth of interest.
This value is compared to the normal effective stress for that depth. The value of
maximum effective stress can be assumed to be constant at 11225 psi and the
value of m can be assumed to equal 2.

σact = σmax * (1-exp((ln Rw - ln Rsh)/m))7.47

Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003
At this point, overburden is determined by previously explained methods and the
normal effective stress is determined.

σnorm = S - pnorm (normal effective stress)

Normal effective stress is compared to actual effective stress. If σact is less than
σnorm, then the difference between these two values represents the increase in
pore pressure above normal. If σact is greater than σnorm, the pore pressure is
normal.

Overall, only two calculations are needed to determine maximum and minimum
magnitudes of pore pressure. Once this is done, the shale water resistivity (Rw)
value is considered constant for the well. The only inputs from the log needed
are shale water resistivity (Rw), observed shale resistivity (Rsh), overburden
(PPG) and TVD depth of interest.
Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003

Drilling Exponent (Dxc) Method

The earliest drilling exponent (Jorden and Shirley), was a normalization of

various drilling parameters to show a relationship between drill rate and
differential pressure. A shale will become denser and less porous with depth,
causing the drill rate to decrease and the exponent to increase. With abnormally
pressured shale, the reverse will be seen.

d exponent = log (ROP/60∗RPM)/log(12∗WOB/10Λ6D)

Where: ROP = Rate of Penetration (ft/hr)

RPM = Rotary Speed (rpm)
WOB = Weight on Bit (lbs)
D = Bit Diameter (inches)

Since differential pressure is dependent upon mud density as well as formation

pressure. The d exponent was corrected to show changes in mud weight. Rehm
and McClendon (1971) proposed the following corrected d exponent (Dxc). To
obtain increased accuracy for Dxc, the equivalent circulating density (ECD) can
replace the mud weight input.

Dxc = dx (Pn/MW)

Where: dx = d exponent
Dxc = corrected d exponent
Pn = normal pore pressure (PPG)
MW = mud weight (PPG)
ECD = equivalent circulating density (PPG)

There are several factors which are not taken into account in calculating
corrected d exponent but must be considered in determining pressures. The
factors are bit wear, bit types, hydraulics, jetting, steerable motors, etc.
However, in normal drilling situations, Dxc works fairly well.

The corrected “d” exponent is plotted continually while drilling in an attempt to

identify deviations from an normal compaction trends which might indicate
changes in pore pressure. Cuttings size and shape, cavings, background gas
levels and the presence or absence of connection gases and trip gases are used
in a qualitative manner to estimate the degree overbalance during drilling.
Changes in relative geothermal temperature gradients affecting the mud entering
and returning from the hole is also monitored.

This method was used to help estimate the pore pressure “while drilling” with
surface logging systems (SLS).
Deepwater Oil Company Post-Well Formation Pressures Analysis
OCS-G 0000 Well No. 1
Deepwater Block 000 1/29/2003
Fracture Gradient Prediction Methods

Fracture pressure is the applied pressure which, when equal to or exceeding the
sum of the formation’s horizontal matrix stress and pore pressure, produces a
structural deformation or fracture. There are two types of fracture pressures
normally seen in oilfield operations: fracture initiation and fracture injection
pressure. Mainly, the knowledge of fracture initiation pressure is important to
well planning to determine casing setting depths and drilling fluids weights
programs.

Fracture initiation pressure is the applied pressure which exceeds the formation
strength (tensile), local borehole hoop stresses, horizontal minimum stress σ3
and pore pressure. This pressure can be significantly higher than injection
pressure. When performing a leak-off test (LOT) at the point where the pressure
becomes non-linear, the local hoop stresses around the borehole are beginning
to yield. The point where this occurs is dependent on the type of rock the LOT is
being performed in, elastic (sands) or plastic (shales/salts). It must however be
stressed that the leak- off test be performed up to the point of fracture (not to a
predetermined point) to be considered in the fracture pressure calculations.

Fracture injection pressure is the applied pressure to open up an existing

fracture. It must exceed horizontal minimum stress σ3 and pore pressure.

A majority of the fracture pressure prediction methods is based on using

Poisson’s ratio with elastic rocks (sands). AGIP’s method recognized differences
between elastic and plastic deformation in rocks, their lithology, and the drilling
fluid permeability. This method will be the primary one used for our services.
These steps are stated below.

Elastic Rocks: Sands/Non Penetrating Drilling Fluids:

F = (2ν / 1- ν)(S - p) + p

Elastic Rocks : Sands/ Penetrating Drilling Fluids:

F = 2ν (S - p) + p

Plastic Rocks: Shales and Salts:

F=S (fracture initiation pressure/LOT is equal to

overburden)

where: ν = Poisson’s ratio (use 0.25 clean sands, 0.28 shaley sands)
F = Fracture pressure
p = Pore pressure
S = Overburden
This page intentionally left blank.
 SPE/lADC

SPE/lADC 18714

A Dual Shale Pore Pressure Detection Technique

by T.M. Bryant,Teleco Oilfield
Services,Inc.
SPE Member

Copyright 19S9,SPE/lADC Drilling Conference

This papar was prepared for presentation et the 1989 SPE/lADC Drilling Conference held in New Orleans, Loublane, February 28-March 3, 1989.

TII18 papar was seleoted for

presentation by a SPE/lAOCProgram Oommittee following review of information contained In an Sb5triICfaubmttted by the author(s).Contente of
as prermted, have not bean reviewed by the International Association of Drilling ContractOreor tfXI SOClOfY of Petroleum Enginaara and are subject to CorrectIonby Ihe
The material, as presented, does not nacaaesrllyrefleot any position of SPE or the IADO, its dfloere, or members. Papera presented at SPE/lADO maatlngaare subject to p
review by Editorial Committees of SPE and the IADC. Permission to copy 18raafrbtad to an abstract of not more than 200 words. Hluatretionsmay not be copbd, The abstr
containconspicuousacknowfettgmentof Mere and by whomthe paperb presented.write Publbstbns Manager,SPE,P.O.Sox822828,Rlchsrclson,TX [email protected],7209S

AMm!2t demonstrate the technique’s consistency

principles, quant”tative accuracy, and rel
A dual-basis application of the shale resisti;::: ease of use.
pore pressure detection technique has
developed that emphasizes the avai1abi1ity of MWD Jntroduct$on
formation resistivitymeasurements while minimizing
input from the operator. Abnormal format”on 9ressures aenerated
compaction-relatedmechanisms have f& many
This method is based upon accepted petrophysical been evaluated through the use of techn
and mechanical principies. relating a measurement of some formation-rela
characteristic to depth. These measure
Operator inputs are streamlined, requiring include gamna radiation, resistivity
one-time-onlyentry of local air gap, water depth, conductivity, sonic transit time, density, se
normal pore pressure gradient, and a baseline value travel time, the drilling exponent, and other
of -shalewater resistivity. well known or used measurements. The measure
vary both in the manner by which they are acq
MWD shale resistivity is utilized in the Archie (e.g., cuttin~s analysis, electrical and acous
equation to solve for shale porosity. The wireline logglng, etc.) and the time when the
Baldwin-Butler normal shale compaction relationship made (e.g., before, during, or after drill
uses this porosity to estimate an actual vertical These variations have an obvious effect on
matrix stress component. Overburden pressure is quality of the measurentent. Differences i
calculated using a generalized Tertiary basin application of any particular measurement ac
equation. The difference between normal and actual for the variet.yyofbetech;::;es by which
matrix stress values represents the magnitude of measurement for evalu
change from normal to abnormal pore pressure compaction-relatedpore pressure.
conditions.
There are numerous pore pressure estim
Unlike traditional empirical approaches and other, techniques associated solely with a form
more recent petrophysical-mechanical techniques, resistivity measurement. Each technique
this technique does not need data froma normally particular advantages and disadvantages, bu
pressured section for establishment of a normal share certain traits. Each is applicable onl
trend, does not require the degree of local the clay-shale litholoqy, and each is based up
knowledge associated with the useof overlays, and expected response with increasing depth
involves neither guesswork nor comprehensive normally pressured shale formations.
expertise on the part of the.operator.
Very early applications of these shale resis
The model has been successfully tested on techniques involved the “equivalent depth” m
approximately 100 well locations in the Gulf of as described in the classic work of Rubey
Mexico, offshore Alaska, North Sea, and Niger Hubbertl. Their theory states that the m
Delta. Field examples are presented that (i.e., framework) pressure of an overpre
.--.----------------.-------------.------------ shale of a given resistivity is considered equ
references and illustrations end of paper magnitude to the matrix pressure of an over
normally pressured shale possessing the same
 A DUAL SHALEPORE PRESSURE DETECTIONTECHNIQUE SPE1IADC 1

resistivity. The magnitude of the abnormal pore technique valid for worldwide use, employing
pressure is the difference between the overburden mininximof user inputs, and requiring minimal use
pressure at the depth of interest and the matrix knowledge and expertise, A further requisite fo
pressure at the equivalent depth. One of the this technique was that it would not involve th
absolute requirements of this technique involves normal parameter trend line which is characterist
the acquisition of shale resistivity data from the of shale resistivity ratio techniques.
normally pressured zone of sufficient quantity to
establish a reliable compaction profile. It is The developed technique utilizes both formati
possible with insufficient compaction data to have ganinaradiation and resistivity. It requires s
an equivalent depth intersectionoccurring outside inputs entered by the operator on a once-on
the geologic section. basis. These are:

Most shale resistivity pore pressure techniques 1) local air gap

utilize ratios of observed (Rb,) and normal (Rn) 2) local water depth
values at a given true vertica’1depth. Larly (mld- 3) normal pore pressure density
60s) techniques, such as those of Hottman and 4) minimum shale ganwnaray value
Johnsonz, Matthews and Kelly3, and Wallace4,were 5 maximum shale gamna ray value
empirical, and typically based upon fewer than two 61 shale water resistivity
dozen wells in the Texas-Louisiana Gulf Coast
Use of these methods is geographicallyand Air gap and water depth are used to adjust t
~~~~~~cally limited. overburden algorithm for local conditions. Th
overburden equation is based upon data obtained b
Foster and Whalens advanced an early petrophysical BelllGfrom the Texas Gulf Coast, and is expressed
and mechanical technique based upon the work of as:
Rubey and Hubbert and ArchieG. Use of this method
requires determination of shale ~ (e.g., using a
wireline SP log) and construction of a shale S/D= E + ( 2.64E-05 * D ) ..(1)
formation factor trend. - ( 1.97E-09 *DA2 )
+ ( 6.60E-14 * 0“3 )
Throughout the late 60s and 70s, numerous - ( 5.97E-19 *OA4)
:;~~;:?:tOrsFe;$!!n’’O’~aneFe;:; a;:cp;;;%
Pilkington and Fer\114;and so on) popularized: where S is the overburden pressure in psi, E
enhanced, and otherwise promoted the use of some the overburden gradient at the mudline (usual
form or another of the early, basic ratio given by the value O.850 psi/ft), and D is the tr
techniques. Without exception, each of these vertical depth, in feet, subsequent to subtracti
authors discussed the major disadvantage associated of both the air gap and the water depth. T
with the shale ratio technique: establishmentof calculated overburden at a given true vertic
the normal resistivity trend line. The problems depth is the sumof this calculated formation
associated with the normal (Rn) trend line include: overburden and the overburden contributed by t
its variable shape, slope and position; the body of water.
frequent requirement for more than one trend, or
for shifts i~f the working trend; t~th~;solute The author’s experience has shown that, wh
necessity obtaining data ‘ the adjusted for local conditions, this equation yiel
near-surface, normally pressured zone; the need in confident, reasonable overburden values, on
some areas for knowledge of formation age; and worldwide basis, for many offshore Tertiary basi
interpretation requiring relatively skilled, which are of an essentially tectonically relax
experienced users. This last disadvantage led many nature. Validity for such encompassing use h
‘oil and service companies to develop pore pressure been confirmed indirectly via the fracture pressu
overlays as a means to obviate the need to have investigationsof Breckels and van Eckelen17, a
such users at the rigsite. However, because of by this author’s comparisons with local densi
their typ:cally empirical and local basis, overlay data.
usage often created worse problems than it was
meant to solve. The equivalent mud density of the normal po
pressure is used in the interval-by-inter
More recently, Holbrook and Hauck15 promoted a calculation of normal pore pressure (pn) and norm
petrophysical-mechanical model based upon the matrix pressure (an), as expressed by Terzaghi’
principles put forth by Terzaghi, Archie, and Rubey relationship:
and Hubbert. Although not a resistivity ratio
technique, it suffers similar disadvantages in that
it requires too many inputs which, by their nature, an =S-pn ..(2)
require a ievel of local knowledge and experience
not possessed by the average user.
The minimum and maximum clay-shale gamna ray valu
define a window used to discriminate the sha
The Techniqu~ Iithology. Since the processes of compactio
diagenesis, groundwater dissolution a
The numerous disadvantages of the shale pore precipitation of secondary minerals, and t
pressure techniques discussed above prompted a generation of abnormal pore pressures may influe
motivation to develop a reliable, theoretical the natural radiation characteristics of a cl
/.
SPE/IAOC 18714 T.M. BRYANT

shale, this wfndow may require adjustments during As with the shale RW situation, there are nume
the course of drilling a well. factors which may influence the magnitude of
porosity exponent. A discussion of this subj
The shale water resistivity value ~ represents a particularly pertaining to clay-shales, is be
variable within Archie’s formation factor equation: the scope of this paper. The author has chose
treat the orosity exponent m as a constant and
suggests tKe value of 2.0 is not unreasona
F = Ro/Rw= I/I#M ..(3) considering the plate-like shape and comparati
high surface area of clays.

There is very little published data regarding the Based upon published data from numerous stud
variation of shale water resistivity (or salinity) Baldwin and Butler21derived a power-law compac
within a sedimentary basin. Few researcherswould relationship for clay shales:
dispute the fact that shale pore water undergoes
both physical and chemical alteration during
compaction and with time. There is, however, Z=z ~x * SA6.35
..(
considerable disagreement concerning the manner by
which shale water resistivity varies with depth.
Foster and Whalen, for instance, suggested that where S, the solidity, is the complement
shale RW simply be considered the same as that of porosity, and Z~,Xthe depth at which S = 1.0.
an adJacent sand. They implied a pattern of
increasing shale water salinity with increasing Using Rubey and Hubbert’s equilibrium d
depth for normally pressured shales. Chilingarian relationship, 8aldwin and 8utler related bu
et al.lg, on the other hand, provided data depth Z to actual matrix pressure U~C~:
indicating that shale water salinity decreased with
increasin overburden pressure. Recently, Morton
and Land”$ described three differing water types u act = C7max
* (z/zmx)~l”16 ..(
within the Oligocene Frio Formation of the Texas
Gulf Coast. Although they concluded that depth-
related salinity changes were of an extremely Substituting (1 - @)AG”35 for (Z/ZmX), ma
variable nature, they found that salinity generally pressure may be related to both shale porosity
increased within the overpressured zone. burial depth:

Owing to the variable nature of shale water

chemistry with depth, it would appear that the oact = umx * (1 - #)~’”35 ..(
selection process for a shale ~ value (for use
with this technique) might be one of random chance.
However, several simplifying assumptions have been Equation 7 may then be adjusted for use with
made to reduce this uncertainty. It is assumed previously described generalized Tertiary b
that shale water salinity increases with depth, overburden expression (equation 1):
from approximately 35,000mg/1 TDS at the surface
(correspondingto normal seawater) to approximately
80,000 mg/1 TDS for a general depth zone of ..(
interest. This zone of interest may be regarded as
equivalent to a depth ran e conforming to a minimum
formation temperature o! 180 ‘F and a maximum The actual matrix pressure aq, is determined
temperature of approximately 400 “F. Considering substituting the right hand STc1
e of equation 4
that many marine Tertiary basins are characterized the shale porosity# in equation 8:
bygeothennal gradients ranging from 1.5 to 2.0
OF/lOO ft, a typical zone of interest may span ..
depths as shallowas 9000 ft (2800 m) and as deep
as 20,000 ft (6096 m). a~ct = amx * (1 - exp( (ln RW - lnR~,)/m))A7-4’
The above assumptions require shale ~ values
ranging from approximately 0.02 to 0.04 ohms. For
the purposes of this technique, shale RW is treated where a~Ct is calculated for each depth
as a constant (within’this range) independent of interest.
depth.
For abnormally pressured formations, as
Both shale water resistivity ~ and corrected MWD pressure increases, the (vertical)matrix pres
shale resistivity R~h are variables in the Archie decreases. This implies that the overbu
equation, rearranged to solve for shale porosity: pressure SO of an overpressured formation at
vertical depth Z is, for all practical purpo
the same as S,, the overburden pressure o
ql=exp( (ln Rw - lnR~h)/m) . . (4) normally pressured formation at the same depth
follows that:

where the porosity exponent m may assume an

average value of 2.0, and exp = e, the base of ao + Po =an+pn ..(10)
Napierian logarithms.
4 A DUAL SHALEPOREPRESSUREDETECTIONTECHNIQUE SPE/ IADC

Given that the change in pore pressure 6P is to 8800 ft. TVD (2073 to 2682m), covering portio
equivalent to the difference between pO and pn, the of both the 12.25 in. (311.21nn) and 8.50 i
increase in pore pressure above normal maY be (215.911MI)hole sections.
determined by subtracting the calculated actual
matrix pressure u~C~ from the calculated normal With the exception of a fifty foot interval betwe
matrix pressure on: 7150 to 7200 ft. (2179 to 2195m), where calculat
pore pressure avera?ed 10.2 ppg (1222 kg/m’), po
pressure was essentially normal (8.7 ppg, or 10
L5p= on - Uact kg/n?) until 7300 ft. (2225 m). At this de t
9
pore pressure increased to 9.9 ppg (1186 kgm3)
remaining at this magnitude until 7650 ft. (23
Table 1 illustrates calculated results of the m). Mud density for this interval was maintain
described technique. at 9.9 ppg, producing a balanced drilling condit
characterized by increasing drilled gas level
Pore pressure rose to 10.5 ppg (1258 kg/m’) at 76
Field Examp1es ft. (2329 m); penetration of a sand at 7680 f
{2341m) necessitated circulation through the ?ho
The technique has been tested on approximately 100 and a mud density increase from 9.9 to 10.5 pp
wells located in the offshore Tertiary basins of Between 7650 and 7850 ft. (2332 to 2393 m), po
the Texas-LouisianaGulf Coast, Alaska, California, pressure was estimated at an average of 10.7 p
the United Kingdom, Norway, Australia, Niger?a, and (1282 kg/n$),with numerous peak valuesat 11.7pp
Trinidad. Criteria for assessment of ‘the (1402 kg/m3). Throughout this interval, backgrou
validation of predicted results have been based gas levels were high, sloughing shale was
upon hole stability observations, kicks, and RFT problem, and mud density was progressively rais
and gas data. It was found that: to 12.0 ppg (1438 kg/m3).

* 70% of the sample group yielded From 7850 to 8150 ft. (2393 to 2484 m), po
exceptionally reasonable pore pressure pressure is estimated to have regressed to betwe
values when tested with a RW value ranging 9.7 and 10.5 ppg (1162 to 1258 kg/n?). With m
between 0.02 and 0.03 ohms. The wells in density at 12.50 ppg (1498 kg/m3), both ho
this group represented each of the areas problems and drilled gas levels diminished. Fr
noted above. this latter depth, pressure increas
reaching 12.7 ppg (1522 &~?) at 8325 ft. (2537
* 20% of the sample group yielded very and 13.0 ppg (1558 kg/m3) at the 8460 ft. (2633
reasonable results when tested with a RW casing depth. Both drilled gas and connection g
value ranging between 0.015 and 0.20 ohmns levels required mud density increa~es to 13.2 p
and an overburden equation E value of 0.80. (1582 kg/m3) at the completion ofthls section.
~];m~~h this e;~~re group was geog~~~~cally
to Mississippi of Ex~le 2: Deepwater Mississippi Fan
Louisiana, not all wells of this area are
included in this group. It is the author’s -f shows a portion of the 12.25 in. ho
belief that the proximity to salt domes may section for a well drilled in 1000 ft. (305 m)
be responsible for the lower ~ value for water. An Rw value of 0.025 ohmns was used as
this portion of the sample group. baseline value.

* 5% of the wells required a shale Rw value Throughout the illustrated interval, mud dens
less than that prescribed by the underlyi~q was maintained at 12.5 ppg (1498 kg/m3) to provi
assumptions of the technique. This group stability for this high angle well. The predic
included only the offshore California wells, 12.4 ppg (1486 k /m3) pore pressure at and bel
and in every case the Rw value was 0.10 7700 ft. (2347m ! was affirmed by the presence
ohmns. connection gas.

* 5%of the test group yielded very poor This example clearly demonstrates the validity
results. These wells were characterized by the technique for both deep water environments a
inadequate data. for situations where data from the norma
pressured zone is not available.
* none of the wells in the test group
indicated that the practical range of shale
Rw actually exceeds 0.03 ohmns. Example3: North Sea

The following field examples typify results of the A portion of the 17.50 in. (444.5mn) hole sect
technique. of a North Sea exploration well is shown in -
3 The geologic column is represented by
Example 1: Offshore Texas Shelf ~;edominant Miocene claystone lithol
interspersedwith occasional thin carbonate be
This vertical exploration well was located in 65 As is characteristic of this region, the shall
ft. (20m) of water. The clay shale lithology was depth, normally pressured claystones possess
defined between the limits of 130 to 180 API, and relatively high formation resisti~ity. Underly
shale water resistivity RW defined as O.O2 ohms. these, overpressure development begins at a v
Ftit shows the MWD pore pressure log from 6800 shallow depth.
SPE/lADC 18714 T.M. BRYANT

Using an ~ value of 0.03 ohnsns, the technique be sufficient to control drilled gas lev~~s~
predicts normal pore pressure conditions existent sloughing tendencies within the shales;
down to 1400 m (4593 ft). Amud density of 1400 also not be so high as to create the conditi
kg/n? (11.7 ppg) was maintained in order to ensure favorable to differential sticking of
hole stability within these reactive claystones. bottomhole assembly. This, in fact, was the c
From this latter depth to 1600 m (5249 ft), pore for the sand at 12652 ft. (3856 m), and
pressure steadily increased, attaining an estimated responsible for sidetracking the original hole.
value of 1440 kg/# (12.0 ppg) at 1590m (5217 ft).
Between this depth and anud density adjustment A two hundred foot thick shale encountered at 12
from 1445 to 1650 kg/n? (12,0 to 13.7 ppg) at 1620m ft. (3917m) exhibited a pore pressure in excess
(5315 ft), pore pressure continued to increase to the mud density, necessitating further dens
1670 kg/nf (13.9 ppg). Underbalanced conditions increases to control the gas levels.
were indicated by very high drilled gas levels
(exceeding 40% methane in air), the presence of
connection and trip gas (in excess of 20%), and Example 6: Offshore California
tight hole problems.
As an example of an exception to the previous c
After peaking at the above value, the pore pressure studies, Fiaure Q shows a pore pressure log for
magnitude regressed within the following forty portion of the 8.50 in. hole section of an offsh
meters to 1380 kg/n? (11.5 ppg). Thereafter, the California well.
pressure increased once more to a value of
approximately * 50 k /n? ( 0.4 ppg) of the mud The mineralogy of these Tertiary shales
density of 1650 kg/n? for the remainder of this frequently diatomaceous, reflected in a resistiv
example. Evidence of this close-to-balance greater than is typical for U.S. Gulf Coast shal
condition was given by high drilled This mineralogy required an RWvalue of 0.10 oh
concentrations (sustained between 30 to 60%) ~~~ for use with the technique, as opposed to the us
the frequent presence of connection gas. range of 0.02 to 0.03 ohnsns.

The pore pressure profile shows a peak

Example 4: Offshore Alaska approximately 11.6 ppg (1390 kg/m3) at 4750 ft.
(1448 m), wherenwd density is 11.8pg (14
F54 illustrates
the pore pressure log fora kg/m3). From this depth to 5900 ft. [17g8 m
portion of the 12.25 in. hole section of a several thin intervals averaging about 10.2
directional development well drilled offshore (1222 kg/m3) interrupt an otherwise norma
Alaska. The shale water resistivity baseline value pressured section. Mud density was increased i
is similar to that used for most applications step-wise fashion over the interval, to 14.1
(0.025 ohmns). (1698 kg/m3) at 5275 ft.(160B m), for the purpo
of providing physical stability in this high-an
Unlike the typical clay shales of the U.S. Gulf hole and to control gas from the fractu
Coast, the shales of this Alaskan example possess a formations.
greater quartz content. As a consequence, shale
resistivity tends to be greater. At 5900 ft. (1798 m), pore pressure increases
approximately 12.6 pp (1510 kg/m3), decreases
Pore pressure remained nearly normal (8.7 ppg, or 11.0 ppg (1318 kg/ l?) by6150 ft. (1875 m),
1024 kg/n?) until 7800 ft. (2377 m), where it ~m~a~s to 12.3 ppg (1474 kg/m3) by 6200 ft.
increased to.9.10 ppg (1090 kg/m3). This pore .
pressure was ma~~:~ned throughout the underlying
thick sand. this sand, pore pressure Concw
increased in the shale to 9.3 ppg (1114 kg/n?).
The pore pressure “spike” at 8200 ft. (2499m) was The field examples presented herein of
most probably anomalous (since it was not convincing evidence testifying to the reliabil
maintained over more than one interval). Mud and validity of this dual shale pore press
density for most of this interval was maintained at detection and prediction technique. It
10.0 ppg (1198 kg/m3), primarily as a consequence successfully provided reasonable quantitative p
of the particularnud type. pressure estimates in both shallow and deep wa
Tertiary sand-shale basins on a worldwide basis.

Example 5: Offshore Nigeria From a practical viewpoint, the technique off

ease of use; user interaction is limited to
wo~sinthe pore pressure log for a portion of start-up entry of only six inputs, none of wh
. . hole section of a near-vertical require any spe~[~~dnowled~e. Unlike other sh
exploration” well drilled offshore on the Niger resistivity Ilttle expertise
Delta. Similar to the previous examples, shale interpretation is nec&sary during its predict
. water resistivity is 0.025 ohmns. operation.

As is characteristic of this area, pore”pressure in Unique to this dual shale pore pressure techni
the clay shales may exceed that of the sands, often is its ability to provide pore pressure estima
by as much as 3.0 ppg (359 kg/m3). The dual shale in the absence of data from the normally pressu
pore pressure technique reliably portrays this zone. Reasonable pressure calculations may,
situation. For such a situation, mud density nust fact, be initiated under any conditions, with
 A DUAL SHALEPORE PRESSUREDETECTIONTECHNXXJE SPE I IADC

calibration necessary. These features make the Meeting of the Society of Petroleum Engineer
technique particularly conducive to real-time MWD of AIME, Houston, Texas, Oct. 1968.
applications.
8. Eaton, B.A., “The Effect of Overburden Stre
Nomenclature on Geopressured Prediction from Well Logs,”
J. Pet. Tech., pp.929-934, Aug. 1972.
D true vertical depth (TVD)
E constant in variable overburden equation, 9. Eaton, B.A., “GraphicalMethod Predicts
psi/ft Geopressured Worldwide,” World Oil, July 1976
F Archie formation factor
m Archie porosity exponent 10. Fertl, W.H. and D.J. Timko, “Overpressure
Po pore pressure in overpressured formation Detection from Wireline Logs,” and “Pitfalls
Pn ~or’~.~nsure in normally pressured Overpressure Prediction,” pts. 3 & 4 of 6 p
series “How Downhole Temperatures, Pressures
RO resistivity of a formation 100% saturated Affect Drilling,” World Oil, 1972.
with water, ohmns
Rsh shale resistivity, ohmns 11. Herring, E.A., “Abnormal Pressures Determin
RW shale water resistivity, ohtnns from Logs,” Petroleum Engineer, Nov. 1973.
s Baldwin-Butler solidity
s overburden pressure 12. Fertl, W.H., “Practical Formation Pressure
so overburden pressure of an overpressured Evaluation from Well Logs,” Petroleum Engine
formation VO1.46, pp.56-70, 1974.
s“ overburden pressure of a normally
pressured formation 13. Lane, R.A. and L.A. McPherson, “A Review of
z true vertical depth GeopressuredEvaluation from Well Logs -
zm% true vertical depth where solidity Louisiana Gulf Coast,” SPE #5033, 49th Annu
equals 1.0 Fall Meeting of the Societyof Petroleum
6p change in pore pressure Engineers of AIME, Houston, Texas, Oct. 1974
# porosity
Uact actual (calculated)matrix pressure 14. Pi.lkington,P.E. and W.H. Fertl, “Howto’Avoi
aImx maxinum matrix pressure Errors in Predicting Geopressures,”World Oi
on normal matrix pressure pp.85-89, June 1977.

15. Holbrook, P.W. andM.L. Hauck, “APetrophysic

Mechanical Math Model for Real-Time Wellsite
1. Rubey, W.W. and M.K. Hubbert, “Role of Fluid Pore Pressure/FractureGradient Prediction,”
Pressure in Mechanics of Overthrust Faulting,” SPE #16666, 62nd Annual Technical Conferenc
Bulletin of the Ge~logical Society of America, and Exhibition of the Society of Petroleum
VO1.70, pp.115-205, Feb. 1959. :~~~eers, pp.163-171, Dallas, Texas, Sept.
.
2. Hottman, C.E. and R.K: Johnson, “Estimation of
Formation Pressures from Log-derived Shale 16. Bell, L.N., “Pressures and Fracture Gradient
Properties,” J. Pet. Tech., pp.717-722, June Atlantic Richfield Drilling Manual, June 196
1965.
17. Breckels, I.M. and H.A.M. van Eckelen,
3. Matthews, W.R. and J. Kelly, “Howto predic~’ “Relationshipbetween Horizontal Stress and
Formation Pressure and Fracture Gradient...from Depth in Sedimentary Basins,” J. Pet. Tech.
:~;ct~;k7and Sonic Logs,” Oil and Gas Journal, pp.2191-2197, Sept. 1982.
. .
18. Terzaghir Charles, “Principles of Soil
4. Wallace, W.E., “Abnormal Subsurface Pressures $lechanics,”Engineering News Record, VO1.95,
Measured from Conductivity or Resistivity 1925.
Logs,” The Log Analyst, VO1.4, Feb. 1965.
19. Chilingarian, G.V., Sawabini, C.T. and H.H.
5. Foster, J.B, and H.E. Whalen, “Estimation of Rieke III, “Comparison between
Formation Pressures from Electrical Surveys, Compressibilities of Sands and Clays,” Journ
Offshore Louisiana,” SPE #1200, 40th Annual of Sedimentary Petrology, VO1.43, pp.529-53
Fall Meeting of the Society of Petroleum 1973.
Engineers of AIME, Denver, Colorado, Oct.
1965. 20. Morton, R.A. and L.S. Land, “Regional
Variations in Formation Water Chemistry, Fr
6. Archie, G.E., “The.ElectricalResistivity Log Formation (Oligocene), Texas Gulf Coast,”
as an Aid in Determining Some Reservoir American Association of Petrolem Geologists
Characteristics,”Trans. AIME, VO1.MG, pp.54- Bulletin, vol. 71, pp. 191-206, Feb. 1987.
62, 1942.
21. Baldwin, B. and C.O. Butler, “Compaction
7. Eaton, B.A., “Fracture Gradient Prediction
_-.._ -._.. Curves,” American Association of Petroleum
Techniques and Their Application in Drilling’, ~~~gists Bulletin , VO1.69, pp.622-626, Apr
Stimulation, and Secondary Recovery .
Operations,” SPE #2163, 43rd Annual Fall
 SPE 18714
Air gap
44.0 fr30t
●

Water depth = 130.0 feet

Min. sh81e MWDGR = 130 API
Max. shale MWD GR = 180 API
Normal pore pressure = 8.70 ppg
Shale Rw = 0.0260 ohmms
------------------------
------------------------

DATA FROM . . . . .,...,

TVD API SNC OBDN MATN MATA PORE MW
11844 148.4 3.20 18.8 10.1 9.1 9.7 11.3
11854 148.4
11864 150.7
11874 148,6
11884 146.9
11894 152.8
11904 152.3
11.914 144.9
1.1924 130.7
11934 146.7
11944 144.1
11954 146.0
11964 152.0
11974 161.4
11984 147.7
11994 146,7
12004 146.2 1.87 18.8 10.1 7.2 11.6 11.3
12014 146.9 1.98 18.8 10.1 7.4 11.5 11.3
12024 137.9 2,91 18.8 10.1 8.7 10.2 11.3
12034 157,1 2.74 18.8 10,1 8.5 10.4 11.3
12044 151.4 2.56 18.8 10.1 8.7 2J.6 11.3
12064 113.9 2.69 18.8 10.1 8.4 10.6 11.3
i2064 143,1 4.41 18.8 10,1 10.0 8.9 11.3
12074 142.6 4.90 18.8 10,1 10.3 8.7 11,3
12084 59.3 2.84 18.8 10.1 8,6 8.7 11,3

Table l-Example Inputs and outputo of the duel shale pore pressure model.
Legend:
TVD: True vertical deplh, ft
APl:MWDgamma ray,APlunita
SNCMWDshorf normalcorracted, ohm-meters
OBDN: Calculated overburden, ppg
MATN: Calculated normal matrix gradient, ppg
MATA: Calculated aotual matrix gradient, ppg
PORE: Calculated pore pressure, ppg
MW Mud density, ppg

““o m <‘.1 , *,0 : “,0 SK ,.!,.., , ~ ~., P.,. P.. *,... <.., , ““D CR <,91> E , “w SW ( . ...0. P... P,**,..,
T 0 .------ -- . . . . ------ 1.6. 200 68,~-----------------~0J~
“ II
8.5
11. d “.. ,,, (.., , ,6 a.a
n.d V,t,ht [,,

Fig. l-MWD pore presssure log, otfshore Texas shelf, Fig. 2-MWD pore pressure log for a well drftled in 1000 ft. (205 m) of water,
Trackl:MWDgamma ray(GR) Gulf of Mexico.
Traok 2: MWD short+rormaf, corraoted formation reslstlvity (SNC)
Track 3: Calculated pore pressure and mud weight.
 SW 18714
““0 c, <A*[,
I
““o WC ,...., ,
.,-------------------- ,... h...”,. (ccl

.Vd ..I. M <se>

,

—
T
-
----
L
t
&.
>.-

.-.-
~i
....,
. . . ..-

.. . . . . . . .. .
.- .,. -
..

-%F!
.. . . . . . .

-------- .. . ... ... ...

n.,
$..

.-A.. ‘:

“.<..,
I

‘. .
,=. .
i .

CSc,
““0 WC , .,..., ,06,5
-----------

8.5

‘,
P.,.

Iivd
P..,,.,,

—
U.s*ht 199
(,.
---------

!-
:-.
,.,

--------
--.-
.-.%
-,...
----
d .
“%
,.-. .. .

I
“w SW (...., , ,08,5 P... P.. *,.,. (.,, ,
-------- . . . . . . . .. -.1.3. H“O Cl <API > 2$ ““o WC ,...., > ,08.3
““* ‘“”’”””” “
0.5 n.4 V.a,
ht(...
> ,$ . . . . . . . . . . . . . . .

*.S ““d ““h’ ‘“’”

T
.L. .
..
Fig, 4-fAWD pore pressure log, offshore Aleska.
::.
--------
.. . . . . . . .
. . . . . . . . .,-.

-.--f,. - ,
.. . . . .. . ----

.-
.
. . .. . . . . . . -.
. .
:..
—
. . . . . .. Iwo CR <WI)
2$
““o WC ( . . . . . , ,08,@ P.,P.....,.
*
--.----------------~s:’
[*

HV4 V,lsht 1...

6.5
I
\ I .,, -------
Fig. S-MWD pore preesure log, offshore Nigeria.

.F\-
,.-.
,*:*

E~
.....
::, ,..
b.

.. . . . . . . . . ,
—...
- I I ----
-- ~.--.f.-
.:.
----
,=: : !-.-.-,-.-;
. . . . . .. . . .
.. . . . . . . . . .

.——L—.—— I

H“D CR ,&P1 , 25 ““D SNC <..... , ,08,5 P., * ?,.,,...

------_------------~s
(

n“d v.,lhn C.
*.5

Fig. 6-MWD pore preeeure log, offshore Cellfornlo.