See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262233382

Estimation of Pore Pressure Gradient and Fracture Gradient from Well Logs: A
theoretical analysis of techniques in use

Conference Paper · September 2009

CITATIONS READS
2 1,089

3 authors, including:

Rima Chatterjee Ashani Kundan

Indian Institute of Technology (ISM) Dhanbad Oil and Natural Gas Corporation Limited,Mumbai,. India
74 PUBLICATIONS   462 CITATIONS    11 PUBLICATIONS   6 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Preparation of Indian Stress Map for Analysis of Crustal Motion and Seismotectonics View project

All content following this page was uploaded by Suman Paul on 07 August 2015.

The user has requested enhancement of the downloaded file.

 Estimation of Pore Pressure Gradient and Fracture Gradient from Well Logs:
A theoretical analysis of techniques in use

Suman Paul*, Rima Chatterjee* and Ashani Kundan**

* Indian School of Mines University, Dhanbad, ** ONGC, Mumbai

E mail id: [email protected]

Abstract

The current paper concerns itself with the theoretical analysis of techniques in use for estimating Pore
Pressure Gradient and Fracture Gradient. The work presented in this paper brings out important insights into
the necessary conditions that must be satisfied by rock properties, for the different techniques to be usable,
and thereby gives physical meaning of the different parameters empirically adjusted, while making
predictions of these gradients and validating them. The study presented indicates that these parameters can
be forward modelled. An example of this is Eaton’s equation, which has been analysed by the authors and
contrasted with Miller and other equations that concern these gradients. The relation between effective
stress and porosity has been analysed and the approach allows for forward modelling the log properties of
mud supported rocks under conditions of normal compaction. Studies such as these help appreciate the
uncertainty and reasons thereof and lead to better appreciation of the confidence level of the predictions.
This in turn is important as both pre-drill as well as syn-drill analyses have significant impact on drilling a
safe and useful well since the projections guide the development of the mud schedule, the casing program, rig
selection and even wellhead ratings. Detailed post-drill analysis of data if integrated with results of the
studies, have the potential to lead to the correct lessons learnt, thereby minimising risk and uncertainty in
future drilling.

Introduction

Pore Pressure Gradient and Fracture Gradient considerations impact the technical merits as well as the
financial aspect of the well plan (Chennakrishnan, 2008). In areas where elevated Pore Pressure Gradients
are known to cause difficulty for drillers, having an accurate pressure prediction at the proposed location is
critical to a successful drilling operation. Pre-drill estimation of Pore Pressure Gradient is the standard
practice for major oil companies. This should come as no surprise, as the reasons are quite compelling. Pore
Pressure Gradient and Fracture Gradient information guides the development of the mud schedule, the casing
program, rig selection and wellhead ratings. Each of these aspects of well planning is capital intensive and
benefit from having a good pre-drill estimate of Pore Pressure Gradient (Mukherjee et al., 2009). In fact, it is
hard to imagine budgeting a multi-million dollar well without this crucial information.

Pore Pressure Gradient analysis can be useful in understanding geological influences on hydrocarbon
accumulation. It is better to drill the flank of a structure rather than its highest point where higher pressure
within the gas cap present more difficult drilling problems. Further, hydrocarbon accumulation favours
slightly lowered Pore Pressure within zones of elevated pressures. Identification of these zones, aids in the
overall exploration of petroleum reserves. Gas, due its buoyancy, can induce abnormally high formation
pressures at very shallow depths. Shallow gas hazards present an important risk while drilling. Pore Pressure
from seismic, together with lithology discrimination, can often identify these zones. The economic impact of
having even one Pore Pressure related event such as a kick or stuck pipe far exceeds the cost of preparing a
detailed analysis of the Pore Pressure for a planned well. Anticipating the pressure ramp and its magnitude
will prepare the driller for what lies ahead. This is far more desirable than having contingency plans to
respond to unscheduled pressure events. It is well known that drilling rates improve as the mud weight
approaches the actual Pore Pressure Gradient. An accurate pressure prediction will aid in achieving this goal
and thus lower the overall cost of drilling the well thorough improved drilling rate. Knowing the depth for the
high pressure allows casing program optimization, increasing the probability of reaching the objective
formations. When geologic uncertainty is reduced a casing string may sometimes be omitted, resulting in
tremendous cost savings. Formation damage and underground blowouts are very costly events. Having a
good pre-drill estimate of Pore Pressure Gradient and Fracture Gradient can significantly decrease the
probability of these serious situations. Pore Pressures can be either normal or abnormal or subnormal.
Normal Pore Pressure will be the hydrostatic pressure due to the average density and vertical depth of the
column of fluids above a particular point in the geological section, that is, to the water table or sea level. The
convention is that abnormal pressures are higher than normal and subnormal pressures are lower.

The main objective of this paper is to analyse (a) the different equations such as: Eaton’s equation, Miller’s
equation, Bowers’s equation and Gassmann’s equation used for estimation of Pore Pressure Gradient and (b)
the different equations such as; Matthew and Kelly’s equation and Pennebaker’s equation used for estimation
of Fracture Gradient.

Methodology

From the well log data, Pore Pressure Gradient and Fracture Gradient have been calculated as described in
the flow chart (Figure 1).

Seismic data (Velocity data) / Log data (LWD / Density log / Porosity log data)

Density (ρ) Calculation

Overburden Gradient (OBG) Calculation

Normal Compaction (NCT) Trend / Compaction Trend (CT) Calculation

Pore Pressure (PP) Calculation

Fracture Gradient (FG) Calculation

Figure 1: Flow Chart for Pore Pressure and Fracture Gradient Estimation

Analysis of Pore Pressure Gradient Equations

Eaton’s Equation and Analysis

The Eaton’s equation (Eaton, 1972 and Eaton, 1975) is,

(1)

where, P = (PP)observed = Observed Pore Pressure

S = P0 = Over burden pressure
Pn = Normal Compaction
Y = Resistivity or Δt etc.
Yn = Resistivity or Δt at normal condition
k = An exponent (Eaton’s exponent)

Analysis:

Eaton’s equation can be written as,

(PP)observed = P0 - (P0 - Pn) (2)

or, P0 - (PP)observed = (P0 - Pn)

or, ,

where, σe = P0 – PP and (σe)n = Effective Stress at normal compaction.

or,

or, , where, Rn = Resistivity at normal compaction. (3)

or, (σe)observed фobservedmk = (σe)n фnmk [As we know, R0 = 1 / фm ] (4)

Let, mk = n, then (σe . фn) is a constant, then σe = C ф–n, where, C = Constant (5)

Eaton’s equation can be written in this form also. This equation relates Effective Stress with Porosity which
indicates Effective Stress increases Porosity decreases.

Miller’s Equation and Analysis

The Miller’s equation (Miller, 1995) is,

V = Vmatrix - (Vmatrix – Vmudline) e – λ (σe)n (6)

where, V = Velocity
Vmatrix = Velocity at matrix (14000 ft / sec to 17000 ft / sec for most shales)
Vmudline = Velocity at mudline (5000 ft / sec)
λ = It is an empirical parameter that yields the best fit for the relation between velocity and effective
stress for the location of interest
(σe)n = Effective Stress assuming normal pore pressure

Analysis:

At mudline, P0 = Water pressure, PP = Water pressure and σe = 0. If we assume the velocity transform plays
V = Vmatrix (1 – ф) + ф Vfluid (7)
or, V = Vmatrix – ф Vmatrix + ф Vfluid
or, Vmatrix - V = ф (Vmatrix - Vfluid)
or, ф = (8)

Now, we can write Miller’s equation as,

Vmatrix - V = (Vmatrix – Vmudline) e – λ (σe)n (9)

or,

or, ф = ф0 – λ (σe)n , where, ф = Porosity and ф0 = Porosity at mudline.

ф = ф0 e – λ (σe)n (10)

Miller’s equation can be written in this form also and this equation is known as Porosity Decline Equation.

Let, assume that a small change in effective stress is Δσe and so, the small change in porosity is Δф. So, we can
write,
= Kф, where Kф = Pore Elastic Constant.

or, Δσe = Kф (11)

As we know the Miller’s equation is,

ф = ф0 e – λ (σe)n (12)
or, Δф = ф0 e – λ (σe)n Δσe – λ
or, Δф = – λ (ф0 e – λ (σe)n) Δσe
or, Δф = – λ ф Δσe [As we know, ф = ф0 e – λ (σe)n]
or, Δσe = -

or, Δσe = Kф [As we know, Δσe = Kф]

Now the questions are, 1) How does Kф behave with respect to applied stress?
and 2) Is Kф independent of applied stress?

If the answer is yes, then ф = ф0 e – λ (σe)n, when λ = (13)

So we can write, (14)

We know that,
Kframe = Kgrain (1 – ф)C

or, = (1 – ф)C

or, = +

or, = +

or, + = +

or, = (15)

If C is independent of ф, then is a constant and ф = ф0 e – λ (σe)n holds.

Bowers’s Equation and Analysis

The Bowers’s equation (Bowers, 1993),

V = Vmudline + A (σe)n B (16)

where, V = Velocity (ft / sec)

Vmudline = Velocity at mudline (≈ 5000 ft / sec)
(σe)n = Effective Stress assuming at normal pressure (psi)
A and B are impirical values that yielded the best fit for the relation between velocity and effective
stress based on the location of where the data was taken.

Analysis:

We know that,
λ= and = . So, we can write, λ = (17)

Here, at mudline Kgrain tends to Kfluid as grains are in suspension and Kfluid is the Bulk Modulus of fluid.
But may be very low, since grains are not in contact with one another in the sense of grains of a rock.

So, when we are at mudline, is low, σe = 0 and Ф = Ф0 (Say). So, any relation for λ to Ф should be of

this form,

λ = A (Ф0 - Ф)ζ +

or, λ = A (Ф0 - Ф)ζ [As is very small and σe = 0] (18)

The basic equation dσe = - translates the equation, dσe = - A (Ф0 - Ф)ζ , when Ф is in such

a range of values such that (Ф0 - Ф)ζ is sufficiently near (Ф0 - Ф)ζ – 1, i.e., Ф is near .
Then, dσe ≈ - A (Ф0 - Ф)ζ – 1 dФ
, where is Constant of Integration. (19)

When Ф = Ф0, σe = 0, then = 0

i.e.,

or, (20)

(21)

Therefore, =

or, = [Let, B = ]

, where is mudline velocity.

Therefore, = , where ( is Constant.

or, = (Say) (22)

or, V = +U

or, V = + , where is a Constant includes both constant U and constant B. (23)

This form leads to Bowers’s Equation. Its unique feature is that σe is related not directly to Ф but to (Ф0 - Ф)
in a power relationship which type of relationships characterises of Eaton’s relationship.

Gassmann’s Equation and Analysis

The Gassmann’s
equation relates the bulk modulus of a rock to its pore, frame and fluid properties. The bulk modulus of a
saturated rock is given by the low frequency Gassmann theory (Gassmann, 1951) as,
 (24)

where, Ksat = Bulk modulus of the saturated rock

Kframe = Bulk modulus of the porous rock frame (drained of any pore fluid)
Kgrain = Bulk modulus of the grain
Kfluid = Bulk modulus of the pore fluid
Ф = Porosity (as fraction)

Analysis:
It is known that is independent of Ф where G is Frame Shear Modulus of Elasticity. Since Gframe can

be written as [Ggrain (1 – Ф)c], where C is independent of Ggrain and Kframe can be written as [Kgrain (1 – Ф)c].
(More of Kgrain and C are discussed later.)

General Considerations - Relation between Effective Stress and Porosity

We know that for saturated frame,

(25)

Let, stress changes to Δσ and let ф changes to Δф. So,

(26)

or, (27)

Now the substitution in the above equation leads to

or,

or, (28)

Gassmann’s equation states that,

From this equation we get,

= + (29)

Let, λ = and also Kframe and Kgrain can be assume to be related through Kframe = Kgrain (1 – ф)C,

where C is a constant and its value was between 1 to 3. Now we can get,

or, =

or, (30)
Hence, Δф =

or, Δф = (31)

Here, λ is in the order of 100’s to 1000’s, so, << 1

or, Δф = (Approx) (32)

We know that, σe = σ - α PP and we make an assumption that Δσe = Δσ. This important assumption means that
Δσe due to Δσ is independent of PP and there by generally that Δσe and Δσ are related through a function not
involving PP. That is to say, the only reason ф change is due to Δσe changes. So, we can write, Δф =
and as differentiates we get,

dф = - dσ.

Now, two cases are shown to show the relation between and ф.

Case – 1: If is independent of ф, then that leads to,

dф = - dσ

or, lnф = - dσ + x, where x is a constant.

or, ф = e - (C / Kgrain) dσe ex, when σe = 0, means only water is present as overburden.

Let, ф = ф0, then ex has to equal ф0. So, ф = ф0 e - (C / Kgrain) dσe (33)
From this we learn that porosity decline constant is actually

Case – 2: If is independent of ф, then that means ‘grain’ of shales depends on ф. Grain to grain

contact elasticity is also heavily dependent on water adsorbed and so Kgrain is dependent on ф by that count
also. C is like grain tortuosity and hence depends upon grain volume (1 - ф) Kgrain should depend on (1 - ф)
and Kgrain also depends on grain volume. For high values of ф, we can assume that is a constant. That

is, from ф0 (Porosity at mudline) down to a value of ф which we shall call фmiller that is a constant. In

ф0 to фmiller region, Kgrain and Cgrain vary together so that is fairly constant. In the region фmiller down to

another value of ф called фeaton (say). We have a different relation as = A ф-ξ. When ф < фeaton, then C
tends to 1 and Kgrain tends to Kmineral, where Kmineral stands for principle mineral of the grain. The relation
= A ф-ξ implies that,

A ф-ξ = dσe

or, A ф-ξ-1 dф = dσe

or, (34)
By plotting log σe versus log ф, we can compute and ξ, that is A, ξ. Now, consider for same overburden

porosity фn for normal compaction case and ф for candidate case (well zone). Then, if S is overburden stress,
P is pressure, Pn is pore pressure at normal compacted case, we have assuming α = 1, so,
(S – P) = σe = .

or, (S – P) = σen =

or,

or, =

or, P = [S – (S - Pn) ] (35)

This leads to the form of Eaton’s equation.

Analysis of Fracture Gradient Equations

Logically Fracture Gradient should mean Formation Breakdown Pressure (FBP) Gradient with depth (i.e.,
Fracture Breakdown Pressure). But Fracture Reopening Pressure (FRP) Gradient what is conveyed by Eaton’s
(1969), Diane’s (1982), and Matthews and Kelly’s (1967) equations. Since, Leak off Pressure (LOP) arises as
recording of existing drilling noticed micro fractures in a number of cases. Fracture Gradient from these
equations is assemble to be compared with LOP. However, for FBP, standard equations have to be used.

In a depositional system, the horizontal stresses σx and σy are expected to be the same, in case the horizontal
stress arises only because of vertical overburden. Let σxe, σye and σze stand for effective stress (i.e., Matrix
Stresses) along x, y and z directions respectively. Let E stand for Young’s Modulus and µ stand for Poisson’s
Ratio.

Let us compute the strain along x direction i.e., Єx (say).

(36)

However, when rock is in geostatic equilibrium Єx should be 0 (same would hold for other directions also).
Imposing the condition Єx = 0 in Equation 41 and the additional condition σxe = σye, and call σxe = σye = σhe, we
get, the condition to be satisfied, to be
0=

or, = (1 - µ)

or, σhe =

If S is overburden, P is Pore Pressure and Biot’s constant α = 1 (assumed), then σze = (S – P). If σh is bulk
horizontal stress, then
σh = (S – P) + P (37)

or, σh = S+ P (38)

In the general case, when α ≠ 1, then σze = (S – α P). If σh is bulk horizontal stress, then
σh = (S – α P) + α P (39)

or, σh = S+α P (40)

Matthews and Kelly’s Equation

If we assume, FRP = Horizontal Matrix Stress + Pore Pressure (P) and also assume that α = 1, then FRP =
(S – P) + P, but denote by Ki, a calibration constant called Stress Ratio (as it converts Vertical

Matrix Stress to Horizontal Matrix Stress), we get FRP = Ki (S – P) + P which is called Matthews and Kelly’s
Equation (Matthews and Kelly, 1967). The Matrix Stress Ratio or also called stress coefficient was made a
variable, as, Matthews and Kelly observed that varied with depth or more accurately with ‘Equivalent

Depth” defined as that depth corresponding to same effective stress as observed formation (i.e., same
property such as ρb or Rt or Interval Velocity). Matthews and Kelly’s stress coefficient is almost constant low
variation with σe down to a depth, and then variation with σe is noticeable and systematic. Below a particular
depth, this systematic exists, but change from that of above mentioned depth ranges (i.e., Effective Stress
ranges).

Pennebaker’s Equation

Pennebaker’s (1968) Equation is similar to Matthews and Kelly’s Equation, but, while Matthews and Kelly’s
Equation investigates and has σe versus Ki correlation, expressed as “Equivalent Depth” versus Ki variation for
forward model. Pennebaker correlates Ki to Depth only (which is unexpected) and thus is, regardless of Pore
Pressure. This correlation of Pennebaker is mainly from South Texas Data. Pennebaker also assumes a
correlation between Depth D and Overburden Stress S. According to Pennebaker, Overburden Gradient
versus Depth trends (Figure 2) belongs to a family of type curves indexed as per the depth of sediments
below mudline, for which the interval transit time is 100 μs / ft. Once the depth is known from Interval
Transit Time versus Depth Data (Figure 3), Overburden Gradient versus Depth Curve is known. At any depth,
we can refer to the selected curve above, to obtain overburden gradient and hence S. Pennebaker next has a
universal type curve, correlating Effective Stress Ratio K (K of Matthews and Kelly’s Correlation) and Depth D
(Figure 4). From this K can be read off for any given D and Horizontal Stress is then, K (S – P) + P.

Figure 2: Overburden Gradient based on average bulk

density between depth point and the surface for different Geologic Age base lines for use in estimating
Fracture Gradients. (Pennebaker, 1968).
Figure 3: Interval Transit Time versus Depth – for use in estimating Fracture Gradients. (Pennebaker, 1968).

Figure 4: Function K versus Depth – for use in estimating Fracture Gradients. (Pennebaker, 1968).

Conclusions

The study presented indicates that these parameters can be forward modeled. An example of this is Eaton’s
equation, which has been analysed by the authors and contrasted with Miller and other equations that
 concern these gradients. The relation between effective stress and porosity has been analysed and the
approach allows for forward modelling the log properties of mud supported rocks under conditions of
normal compaction. Studies such as these help appreciate the uncertainty and reasons thereof and lead to
better appreciation of the confidence level of the predictions. This in turn is important as both pre-drill as
well as syn-drill analyses have significant impact on drilling a safe and useful well since the projections guide
the development of the mud schedule, the casing program, rig selection and even wellhead ratings. Detailed
post-drill analysis of data if integrated with results of the studies, have the potential to lead to the correct
lessons learnt, thereby minimising risk and uncertainty in future drilling.

Acknowledgement

We would like to convey our sincere thanks to Oil and Natural Gas Corporation Limited, Mumbai for
providing us the infrastructural facilities.

We would like to pay our special thanks to Mr. K. M. Sundaram, General Manager (Wells) of Oil and Natural
Gas Corporation Limited, Mumbai for sharing his valuable experiences and guidance to write this paper.

References

Bowers, G. L., 1993, Method for Estimating Pore Fluid Pressure, United States Patent, Patent # 5,200,929,
April 6.

Chennakrishnan, B., 2008, Pore pressure and Wellbore Stability Analysis of CB-ONN-2002/2 Block in
Cambay Basin presented in GEOINDIA in New Delhi, India, September.

Daines, S. R., 1982, Prediction Of Fracture Pressures For Wildcat Wells, Journal of Petroleum Technology,
April, 863 – 872.

Eaton, B. A., 1969, Fracture Gradient Prediction And its Application In Oilfield Operations, Journal of
Petroleum Technology, October, 1353 – 1360.

Eaton, B. A., 1972, The effect of overburden stress on geopressure prediction from well logs, Journal
Petroleum Technology, v. 24, 929 – 934.

Eaton’s B. A., 1975, The Equation For Geopressure Prediction From Well Logs, SPE 50th Annual Fall Meeting,
Dallas, TX, September 28 – October 1, SPE Paper # 5544, 11.

Gassmann, F., 1951, Uber die elastizität poröser medien: Vierteljahrsschrift der Natur forschenden
Gesellschaft in Zürich, v. 96, 1-23.

Matthews, W. R. and Kelly, J., 1967, How to Predict Formation Pressure and Fracture Gradient, Oil and Gas
Journal, February 20, v. 65, 92 – 1066.

Miller, T. W., 1995, New Insights on Natural Hydraulic Fractures Induced By Abnormally High Pressure,
AAPG Bulletin, v. 79, no. 7, 1005 – 1018.

Mukherjee, B. K., Goswami, P. S., Dutta R. and Gauma, M. S., 2009, Evaluation of pressure system in the
fields of Deep offshore-KG Basin, Petrotech, Jan. 11-15, New Delhi, India, No. P09-608.

Pennebaker, E. S., 1968, Seismic data indicate depth, magnitude of abnormal pressures, World Oil., June, 73-
78.

View publication stats