URTeC: 3860563

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Quantifying Variability in Hydraulic Fracture Geometry Using Stage-
by-Stage ISIP Analysis in Geomechanically Heterogeneous Rocks

Erfan Sarvaramini, PETRONAS Energy Canada Ltd., Calgary, AB, Canada

Copyright 2023, Unconventional Resources Technology Conference (URTeC) DOI 10.15530/urtec-2023-3860563

This paper was prepared for presentation at the Unconventional Resources Technology Conference held in Denver, Colorado, USA,
13-15 June 2023.

The URTeC Technical Program Committee accepted this presentation on the basis of information contained in an abstract submitted
by the author(s). The contents of this paper have not been reviewed by URTeC and URTeC does not warrant the accuracy, reliability,
or timeliness of any information herein. All information is the responsibility of, and, is subject to corrections by the author(s). Any
person or entity that relies on any information obtained from this paper does so at their own risk. The information herein does not
necessarily reflect any position of URTeC. Any reproduction, distribution, or storage of any part of this paper by anyone other than the
author without the written consent of URTeC is prohibited.

Abstract
This work presents a novel framework to quantify the variability in hydraulic fracture geometry in a multi-
cluster multi-stage hydraulic fracturing treatment well using stage-by-stage Instantaneous Shut-in Pressure
(ISIP) analysis. The fundamental Sneddon's equation is used to develop a mathematical solution that
accounts for poroelastic stress shadow effects and lateral geomechanical heterogeneity, which leads to non-
uniformity in hydraulic fracture height and length distribution along the horizontal well. Additionally, a
separate solution is proposed for calculating the hydraulic fracture half-length in a multi-stage multi-cluster
horizontal well using the fundamental solution of the Perkins, Kern, and Nordgren (PKN) fracture. It will
be demonstrated that this computationally efficient approach can reasonably replicate the stage-by-stage
hydraulic fracture geometry with minimal cost and is in good agreement with microseismic monitoring and
fracture modeling results.

Introduction
Hydraulic fracturing (HF) is a widely used technique in the oil and gas industry to enhance production from
tight hydrocarbon-bearing formations, such as coal beds, shales, mudstones, and tight sandstones. This
process involves pumping a highly pressurized fluid mixed with sand and chemicals into the well at high
rates to create permeable conduits in the rock (Economides and Nolte, 2000). To ensure the success of this
process, it is crucial to understand the hydraulic fracture geometry, including their half-length and height.
This information allows operators to optimize hydraulic fracturing treatment design and achieve the desired
production results.
Accurately predicting and modeling hydraulic fracture geometry can be challenging due to several factors,
such as complex stress regimes in subtly different geological and geomechanical environments as well as
subsequent fracture interactions in multi-cluster, multi-stage fracture jobs created from different benches
of a multi-well pad (Warpinski, 2008; Mayerhofer, 2011). Hydraulic fracturing diagnostics are widely used
to gather crucial data during the fracturing process to optimize well productivity. The most commonly
employed fracturing diagnostics include microseismic monitoring, fiber optics, tracers, pressure, and rate
transient analysis. While microseismic and fiber optics diagnostics offer in-depth insights into the geometry
of the fracture, including its shape, size, and orientation, they require specialized equipment, expertise, and
URTeC 3860563 2

are expensive to execute. Despite these challenges, combining information obtained from these diagnostics
with advanced fracture modeling techniques and data analysis can improve future fracturing operations
(Cipolla et al. 2010).

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Hydraulic fracturing is a complex and expensive process, and the cost of fracturing diagnostics can add to
the overall operational expense. Operators may choose to run fracturing diagnostics on selected wells within
a pad, often those considered "science-focused" wells. Conducting expensive fracturing diagnostics on
every well within a pad is not always feasible due to logistical and financial considerations. In such cases,
less expensive datasets, or diagnostics, such as pressure and rate transient analysis, flowback analysis or
interference testing can serve as a cost-effective alternative to obtain insights into fracturing behavior.
However, a drawback of these methods is that they often do not provide detailed information on stage-by-
stage fracture geometry, but instead offer an overall view of the fracture geometry in the well (Cipolla et
al. 2000)
The Instantaneous Shut-in Pressure (ISIP) dataset is a valuable tool that can be utilized at the fracturing
stage level and is readily available for every well. ISIP has been commonly used in drilling and completion
operations as a diagnostic tool to assess the efficiency of the fracturing process and the condition of the
wellbore such as potential leaks (Warpinski et al., 1995). ISIP is typically measured once the pumps are
shutoff following each stage of the fracture stimulation, during which the flowing pressure is expected to
rapidly decrease due to frictional dissipation. This is then followed by an excitement period, which results
in a vibration decay of the pressure prior to stabilizing at the ISIP pressure. This period is commonly
referred to as the water hammer period (Montgomery and Smith, 2014).
The ISIP represents average fracture pressure in the intermediate or far-field region away from the fracture
inlet that can be used to estimate fracture height. Nolte proposed a relationship for a PKN (Perkins, Kern,
and Nordgren) fracture, which assumes that the fracture height is proportional to the ratio of fracture
toughness 𝐾𝐼𝑐 and fracture net pressure 𝑝𝑛𝑒𝑡 , i.e., (ℎ ∝ (𝐾𝐼𝑐 /𝑝𝑛𝑒𝑡 )2, where 𝑝𝑛𝑒𝑡 = 𝐼𝑆𝐼𝑃 − 𝜎𝑚𝑖𝑛 represents
the relationship between fracture pressure (ISIP) and closure stress (𝜎𝑚𝑖𝑛 ) (Nolte, 1991; Sarvaramini and
Garagash, 2015). Nolte’s equation can be used to estimate the fracture height for a single fracture if the
fracture and reservoir geomechanical properties are known.
During a multi-stage hydraulic fracturing operation, the fracturing process typically commences at the well
toe, and hydraulic fractures are created sequentially in stages along the lateral well. The interaction between
the fractures generated in each stage can cause stress shadowing effects, which may result in an increase in
fracture pressure. While Nolte's equation can still be used to estimate fracture height, it is unclear whether
this relationship holds in the context of multi-stage hydraulic fracturing, as it falls outside the range of its
assumed conditions. Furthermore, fractures can become tougher as they propagate, making it even more
challenging to estimate 𝐾𝐼𝑐 (Garagash, 2019).
Sneddon (1946) proposed a mathematical model that describes the stress perturbation around a pressurized
fracture, which decays as a function of distance relative to the fracture height. Roussel (2017) developed a
mathematical model based on Sneddon's fundamental solution to quantify stress shadowing effects in multi-
stage hydraulically fractured horizontal wells. The model shows that stress escalation magnitude increases
exponentially as more fractures are added along the well until it reaches a plateau, beyond which the
escalation growth rate becomes slow. By fitting an exponential function to the ISIP data points along the
lateral, an average fracture height can be estimated. However, this work assumed a uniform fracture height,
which may not be valid when geomechanical heterogeneity is significant. In reality, geomechanical
properties, resulting from the intersection of the well with different geological formations and/or variations
in depositional areas, cannot be ignored. This heterogeneity can cause variability in hydraulic fracturing
height and length, leading to deviations from the ideal expected exponential shape of the ISIP data.
Moreover, parent-child or fault-hydraulic fracture interactions can suppress fracture height due to excessive
leak-off, resulting in the non-uniform fracture geometry. Fracture height variability has been linked to ISIP
URTeC 3860563 3

data and microseismic monitoring (Perez, 2021), and attempts have been made to quantify it using an
iterative approach proposed by Moradi (2021).
This work aims to expand on Roussel's previous work by proposing a general solution to the stress

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
escalation theory. The proposed solution will estimate the variability in stage-by-stage fracture height
resulting from lateral geomechanical heterogeneity in the rocks and/or depletion or fault-fracture interaction
effects. This is achieved by using a diverse range of data sets, including construction of a rigorous
Mechanical Earth Model (MEM) that incorporates various geological, geomechanical data derived from
logs calibrated with static properties as well as drilling derived properties, pore pressure and stress
measurements from Diagnostic Fracture Injection Test (DFIT). Once the fracture height is known, it is
straightforward to estimate the fracture half-length. Additionally, a separate mathematical solution is
proposed for calculating the hydraulic fracture half-length in a multi-stage multi-cluster horizontal well
using the fundamental solution of the PKN model.
The application of this method is demonstrated by using a crosscut well deliberately drilled through multiple
geomechanically heterogeneous zones in the Montney Formation in Northeast British Columbia, Canada
and monitored using a downhole microseismic program. It will be illustrated that this computationally fast
approach can replicate the average geometry of a stage-by-stage hydraulic fracture stimulation at relatively
no cost, while being in good agreement with microseismic monitoring and fracture modeling.

1. Quantification of the fracture height

1.1 Sneddon’s stress perturbation theory

Consider a PKN fracture in a rock layer with height h, length 2xf and opening width w, sandwiched between
two impermeable layers, as depicted in Figure 1. This fracture is loaded internally by a fluid pressure pf
resulting from a fluid injected at the rate Qo. The far field in-situ stress field is characterized by three
principal stress directions, i.e., minimum in situ stress (𝜎𝑚𝑖𝑛 ), maximum principal stress (𝜎𝑚𝑎𝑥 ), and
overburden stress (𝜎𝑉 ). The magnitude of the induced stress in the direction of the minimum, maximum and
intermediate confining stresses is given by 𝛥𝜎𝑥𝑥 , 𝛥𝜎𝑦𝑦 , and 𝛥𝜎𝑧𝑧 , respectively. In this paper, the stress
distribution over the fracture plane in the y-z plane is divided into two categories. The stresses that are
distributed within the fracture plane (i.e., 𝜎𝑦𝑦 , and 𝜎𝑧𝑧 ) are known as in-plane stresses, while the stress that
is perpendicular to the fracture plane in the y-z plane (i.e., 𝛥𝜎𝑥𝑥 ,) is known as off-plane stress (Figure 1).

Figure 1: Propagation of a PKN fracture with length 2xf and height h (h ≪ xf.). 𝜎𝑚𝑖𝑛 , 𝜎𝑚𝑎𝑥 and 𝜎𝑉 are the in-situ principal stresses, in the
direction of x, y, and z, respectively. 𝛥𝜎𝑥𝑥 , 𝛥𝜎𝑦𝑦 and 𝛥𝜎𝑧𝑧 are the induced stresses due to the injection of a pressurized fluid.
URTeC 3860563 4

Assuming that the fracture height h is much smaller than the fracture half-length xf (i.e., h ≪ xf ), the
deformation field in any vertical cross section not immediately adjacent to the fracture front can be
approximated as plane-strain. Note that in the Western Canadian Sedimentary Basin (WCSB), where the

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Montney Formation is located, the tectonic stress regime is predominantly in a strike-slip condition, with
the maximum horizontal stress being larger than the overburden stress (i.e., the intermediate stress is the
overburden stress).
Sneddon (1946) proposed a mathematical solution that describes the relationship between the fluid pressure
inside a PKN hydraulic fracture and the resulting stress perturbation on the side parts of the fracture, which
decays as a function of distance relative to the fracture height. The stress perturbation around a PKN fracture
in the x direction (y=0 and z=0) can be expressed mathematically as (Sneddon 1946; Roussel, 2017):
𝛥𝜎𝑥𝑥 1
ℱ𝑥𝑥 (𝜒) = = (1 − 3 ), (1)
𝑝𝑛𝑒𝑡
(1+𝜒2 )2

𝛥𝜎𝑦𝑦 1+𝜒2
ℱ𝑦𝑦 (𝜒) = = 2𝜐 (1 − 3 ), (2)
𝑝𝑛𝑒𝑡
(1+𝜒2 )2

𝛥𝜎𝑧𝑧 1+2𝜒2
ℱ𝑧𝑧 (𝜒) = = (1 − 3 ), (3)
𝑝𝑛𝑒𝑡
(1+𝜒2 )2

where 𝜒 = ℎ/2𝑥, 𝑝𝑛𝑒𝑡 is the fracture net pressure, 𝜐 is the Poisson’ ratio, and ℱ is the normalized stress
perturbation function, which will be referred to as the influence function from here on.
Figure 2 illustrates the influence function ℱ𝑖𝑖 (𝑖 = 𝑥𝑥, 𝑦𝑦, 𝑧𝑧) versus h/2x. A typical value for the Poisson’s
ratio is assumed (in this example, 𝜐 = 0.25). As shown, the in-plane components of the stress field
(𝑖. 𝑒. , 𝛥𝜎𝑦𝑦 , 𝛥𝜎𝑧𝑧 ) are expected to decay more rapidly away from the fracture compared to the off-plane stress
component (𝑖. 𝑒. , 𝛥𝜎𝑥𝑥 ).

Figure 2: Influence function ℱ𝑖𝑖 (𝑖 = 𝑥𝑥, 𝑦𝑦, 𝑧𝑧) versus h/2x

1.2 Stress shadowing in the multi-stage hydraulic fracturing

1.2.1 Geomechanically homogenous rock (uniform fracture height)
Consider a horizontal well that is hydraulic fractured with n equally spaced stages at a distance of Ls
between each stage as illustrated in Figure 3. We assume that the well is placed in rock with similar
URTeC 3860563 5

geomechanical properties, and that the height of each fracture is the same and equal to h. Note that the x-
direction in Figure 3 is considered the direction of the horizontal wellbore.
The injection of a highly pressurized fluid creates a high stress zone around the newly created fracture as

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
described by Sneddon (1946). When a horizontal well is fractured, multiple fractures are created
consecutively, and the stress field in the rock formation can be affected by the previously created fractures.
This phenomenon is known as “stress shadowing” effects, and it can be quantified using Sneddon's equation
(Eq. 1-3), based on a superposition method for multi-stage hydraulic fracturing proposed by Roussel (2017).

Figure 3: A multi-stage horizontal well with the spacing Ls placed in a rock with similar geomechanical properties. The fracture height is h.

1.2.1.1 Off-plane stress escalation during multi-stage hydraulic fracturing

Roussel (2017) developed a mathematical model to calculate the off-plane stress increase (i.e., 𝛥𝜎𝑥𝑥 ) along
a multi-stage hydraulically fractured horizontal well. This model assumes a uniform fracture height along
the well trajectory, which corresponds to a scenario where geomechanical heterogeneity is negligible. This
could occur, for instance, if the horizontal well is placed in rock with similar geomechanical properties.
In view of Sneddon’s Equations (Eq. 1-3), the stress escalation in a multi-stage hydraulic fracturing job is
approximated by (Roussel, 2017)
𝛥𝜎𝑥𝑥 i
= ℱ𝑥𝑥 (χ𝑠 ) (1 + ∑ni=1(ℱ𝑥𝑥 (χ𝑠 )) ), (4)
𝑝𝑛𝑒𝑡

where n is the stage number, χ𝑠 = ℎ/2𝐿𝑠 expressed as a function of the stage spacing 𝐿𝑠 and fracture height
h. Note that the fracture net pressure pnet was a priori unknown in Roussel’s work. However, an exponential
function was fit to a set of ISIP data points along the multi-stage fractured well to obtain an average pnet as
a byproduct of the proposed mathematical solution. In practical applications, pnet can be assumed as a prior
known value, calculated as by 𝑝𝑛𝑒𝑡 = 𝐼𝑆𝐼𝑃𝑒𝑓𝑓 − 𝜎𝑚𝑖𝑛 , where 𝐼𝑆𝐼𝑃𝑒𝑓𝑓 and 𝜎𝑚𝑖𝑛 represent the effective ISIP
and minimum horizontal stress, respectively. These parameters can typically be determined through a
Diagnostic Fracture Injection Test (DFIT). The latter is demonstrated through the measurement of stress
shadowing effects during multi-cluster multi-stage hydraulic fracturing using downhole gauges and optical
fiber (Haustveit et al., 2022).

Figure 4 illustrates an example of stress escalation using Roussel's model (Eq. 4) for a multi-stage hydraulic
fracturing operation with thirty equally spaced stages (n=30) at a given stage spacing Ls. In this example,
the stage spacing is assumed to be equal to half of the fracture height (i.e., χ𝑠 = ℎ/2𝐿𝑠 = 1). Figure 4 shows
URTeC 3860563 6

that the magnitude of stress escalation increases exponentially until it reaches a plateau, beyond which the
escalation growth rate becomes very slow. The magnitude of stress influence function plateau is ≈ 1.8.

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Figure 4: Stress escalation using Roussel’s model for a multi-stage hydraulic fracturing operation with thirty equally spaced stages (n=30) at a
given stage spacing Ls

1.2.1.2 The general solution of off-plane stress plateau

The asymptotic solution for the stress escalation plateau can be obtained from Roussel's model (Eq. 4) by
taking the limit as 𝑛 → ∞. The resulting solution is:

𝛥𝜎𝑥𝑥 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 𝛥𝜎𝑥𝑥 (𝑛) i ℱ𝑥𝑥 (χ𝑠 )

= 𝑙𝑖𝑚 = lim (ℱ𝑥𝑥 (χ𝑠 ) (1 + ∑ni=1(ℱ𝑥𝑥 (χ𝑠 )) )) = . (5)
𝑝𝑛𝑒𝑡 𝑛→∞ 𝑝𝑛𝑒𝑡 n→∞ 1−ℱ𝑥𝑥 (χ𝑠 )

This solution can be modified for multi-cluster fracture designs by substituting χ𝑠 → χ̃𝑠 , where
2𝑛𝑐 ℎ
χ̃𝑠 = ( ), (6)
𝑛𝑐 +1 2𝐿𝑠

where 𝑛𝑐 is the number of entry points per stage.

Substituting Eq.1 in Eq. 5 in yields

𝛥𝜎𝑥𝑥 𝑃𝑙𝑎𝑡𝑒𝑎𝑢
= (1 + χ̃2𝑠 )3/2 − 1. (7)
𝑝𝑛𝑒𝑡

1.2.1.3 Asymptotic solutions of off-plane stress plateau

Mathematically, the induced stress plateau can be described by a general analytical solution:

𝛥𝜎𝑗𝑗 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 ℎ 𝛼
= 𝛽(𝑛𝑐 ) ( ) , ( 𝑗 = 𝑥𝑥, 𝑦𝑦, zz), (8)
𝑝𝑛𝑒𝑡 𝐿𝑠

where 𝛼 and 𝛽 are constants. The value of 𝛼 depends on the ratio ℎ/𝐿𝑠 , and the value of 𝛽 depends on the
number of entry points per stage (𝑛𝑐 ).

In view of Eq. 8, the asymptotic solutions of the limiting cases of the off-plane stress plateau can be obtained
from Eq. 7 and are as follows:
URTeC 3860563 7

𝛥𝜎𝑥𝑥 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 3 𝑛𝑐 2 ℎ 2
= ( ) ( ) ; ℎ ≪ 𝐿𝑠 , (9)
𝑝𝑛𝑒𝑡 2 𝑛𝑐 +1 𝐿𝑠

and

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
𝛥𝜎𝑥𝑥 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 𝑛𝑐 3 ℎ 3
=( ) ( ) ; ℎ ≫ 𝐿𝑠 , (10)
𝑝𝑛𝑒𝑡 𝑛𝑐 +1 𝐿𝑠

respectively. It should be noted that the exponent 𝛼 in the general analytical solution of the off-plane
induced stress plateau (Eq. 8) varies between 2 ≤ 𝛼 ≤ 3. This means that the off-plane stress can
significantly increase when the average fracture height becomes greater than the stage spacing length (h>
Ls).

Figure 5 shows how the off-plane stress plateau varies as a function of h/Ls for various cluster designs
(Roussel, 2017). This graph also includes the asymptotic solutions (Eq. 9-10) for comparison in red dashed
line.

Figure 5: Off-plane stress plateau as a function of h/Ls for various cluster designs (Roussel, 2017). This graph also includes the asymptotic
solutions for comparison in red dashed lines (Eq. 9-10)

1.2.1.4 In-plane stress escalation during multi-stage hydraulic fracturing

The formulation of induced stress in in-plane components can be done similarly to the off-plane stress. The
stress escalation equation (Eq. 4) can be written in general form as:

𝛥𝜎𝑗𝑗 i
= ℱ𝑗𝑗 (χ𝑠 ) (1 + ∑ni=1 (ℱ𝑗𝑗 (χ𝑠 )) ) , ( 𝑗 = 𝑦𝑦, 𝑧𝑧). (11)
𝑝𝑛𝑒𝑡

Figure 6 shows an example of stress escalation for 𝛥𝜎𝑗𝑗 (𝑗 = 𝑥𝑥, 𝑦𝑦, 𝑧𝑧) as a function of the stage number
assuming χ𝑠 = ℎ/2𝐿𝑠 = 1. As illustrated, the magnitude of induced stresses in the in-plane yy and zz
directions is significantly lower than that of off-plane xx direction. Notably, in this example, 𝛥𝜎𝑧𝑧 undergoes
stress de-escalation.
URTeC 3860563 8

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Figure 6: An example of stress escalation for 𝛥𝜎𝑗𝑗 (𝑗 = 𝑥𝑥, 𝑦𝑦, 𝑧𝑧) as a function of the stage number when χ𝑠 = ℎ/2𝐿𝑠 = 1

1.2.1.5 The general solution of the in-plane stress plateau

A similar approach can be used to formulate the in-plane stress components as for the off-plane stress. In
this case, the induced stress plateau for the in-plane stress components can be expressed as:
3
𝛥𝜎𝑦𝑦 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 (1+χ̃2 )2 −(1+2χ̃2 )
= , (12)
𝑝𝑛𝑒𝑡 1+2χ̃2

and
3
2𝜐((1+χ̃2 )2 −1−χ̃2 )
𝛥𝜎𝑧𝑧 𝑃𝑙𝑎𝑡𝑒𝑎𝑢
= 3 3 . (13)
𝑝𝑛𝑒𝑡
(1+χ̃2 )2 −2𝜐((1+χ̃2 )2 −1−χ̃2 )

1.2.1.6 Asymptotic solutions for the in-plane stress plateau

The limiting cases of the in-plane stress plateau have the following asymptotic solutions:

𝛥𝜎𝑦𝑦 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 1 𝑛𝑐 2 ℎ 2
=− ( ) ( ) ; (ℎ ≪ 𝐿𝑠 ), (14)
𝑝𝑛𝑒𝑡 2 𝑛𝑐 +1 𝐿𝑠

𝛥𝜎𝑦𝑦 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 𝑛𝑐 ℎ
= 2( ) ( ); (ℎ ≫ 𝐿𝑠 ), (15)
𝑝𝑛𝑒𝑡 𝑛𝑐 +1 𝐿𝑠

and

𝛥𝜎𝑧𝑧 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 𝑛𝑐 2 ℎ 2
= 𝜐( ) ( ) ; (ℎ ≪ 𝐿𝑠 ), (16)
𝑝𝑛𝑒𝑡 𝑛𝑐 +1 𝐿𝑠

𝛥𝜎𝑧𝑧 𝑃𝑙𝑎𝑡𝑒𝑎𝑢 2𝜐
=( ) ; (ℎ ≫ 𝐿𝑠 ), (17)
𝑝𝑛𝑒𝑡 1−2𝜐

respectively.
URTeC 3860563 9

1.2.2 Geomechanically heterogeneous rock (non-uniform fracture height)

In the previous section, we reviewed the induced stress escalation formulation, assuming that any

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
geomechanical heterogeneities along the wellbore trajectory were negligible. However, in practical
scenarios, lateral geomechanical properties cannot be ignored, and they mainly occur due to two factors:
(1) the intersection of the horizontal well with different geological formations, and (2) inherent lateral
geomechanical heterogeneity caused by variations in geological depositional areas. The presence of lateral
geomechanical heterogeneity can result in variability in the hydraulic fracturing height and length.

In the following section, we will provide a general solution to quantify the fracture geometry in the presence
of lateral heterogeneity along a horizontal well. Let us consider a situation where a horizontal well has been
drilled through multiple geological horizons with varying geomechanical properties, as illustrated in Figure
7. Suppose a hydraulic fracture design has been proposed with n single cluster hydraulic fracturing stages
placed along the lateral at the spacing Lsi (i = 1, …, n).

Figure 7: A multi-stage horizontal well with the spacing Lsi (i = 1, …, n-1) crosscutting several geological horizons with varying geomechanical
properties

We assume that each fracture stage has encountered a rock formation with a net fracture pressure, denoted
(𝑖) (𝑖)
as 𝑝𝑛𝑒𝑡 = 𝐼𝑆𝐼𝑃′ − 𝜎 𝑖 𝑚𝑖𝑛 , , where i ranges from 1 to n. This pressure is determined by the difference
between the virgin 𝐼𝑆𝐼𝑃′ (which is not influenced by the shadowing effects) and the minimum confining
(𝑛)
stress (𝜎𝑚𝑖𝑛 ). The net fracture pressure, denoted as 𝑝𝑛𝑒𝑡 , is directly proportional to the fracture toughness
(Nolte, 1991) or Unconfined Compressive Strength (UCS) of the rock formation. Higher values of fracture
toughness and UCS indicate that the rock has a greater ability to resist fracture propagation, which requires
a higher net fracture pressure to induce fractures.

In practice, measuring fracture toughness in a laboratory setting is not always feasible, and therefore, the
UCS can be used as a substitute to inform the values of pnet. UCS can be calculated through various methods,
including seismic amplitude variation with offset (AVO) inversion, or using geomechanical properties
derived from drilling data, which have been calibrated against core laboratory measurements. These
methods provide a reliable estimate of rock strength, even in the absence of direct laboratory measurements,
and can be useful in assessing the mechanical behavior of in-situ rocks. It is important to note that in this
study, we assume that the pnet is directly proportional to the UCS. This relationship assumes that the rock's
ability to resist fracture propagation is primarily governed by its UCS value.
URTeC 3860563 10

In a multi-stage hydraulic fracturing job, the fracturing process typically begins at the well toe, where the
initial shut-in pressure (ISIP) is equal to the virgin ISIP' of the formation:
(1)
𝐼𝑆𝐼𝑃 (1) = 𝐼𝑆𝐼𝑃′

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
, (18)

The placement of the first hydraulic fracturing stage generates induced stresses on the broadside of the
fracture, which deteriorates with increasing distance relative to the fracture height (proportional to x/h), as
explained by Sneddon's equation (Eq. 1-3). This creates a traction load, 𝑻(1), on the surface of the second
fracture within the second stage. The increase in the ISIP resulting from the traction load 𝑻(1) applied by
the first stage on the second stage can be mathematically represented as
(2) (1)
𝐼𝑆𝐼𝑃 (2) = 𝐼𝑆𝐼𝑃′ + 𝑻(1) ; 𝑻(1) =ℱ𝑥𝑥 (χ1 ) 𝑝𝑛𝑒𝑡 , (19)

where
1 ℎ1
ℱ𝑥𝑥 (𝜒1 ) = 1 − (1+𝜒 2 3/2 ; where 𝜒1 = . (20)
1 ) (𝐿𝑠 1 +𝐿𝑠 2 )

A similar process occurs in subsequent hydraulic fracturing stages, where the traction forces accumulate
from the previous fractures, leading to a gradual increase in ISIP. This increase can be mathematically
represented for the subsequent stages as:

Stage 3:
(3) (2)
𝐼𝑆𝐼𝑃 (3) = 𝐼𝑆𝐼𝑃′ + 𝑻(2) ; 𝑻(2) =ℱ𝑥𝑥 (χ2 )(𝑝𝑛𝑒𝑡 + 𝑻(1) ). (21)

Stage 4:
(4) (3)
𝐼𝑆𝐼𝑃 (4) = 𝐼𝑆𝐼𝑃′ + 𝑻(3) ; 𝑻(3) =ℱ𝑥𝑥 (χ3 )(𝑝𝑛𝑒𝑡 + 𝑻(2) ). (22)

Stage n-1:
(𝑛−1) (𝑛−2)
𝐼𝑆𝐼𝑃 (𝑛−1) = 𝐼𝑆𝐼𝑃′ + 𝑻(𝑛−2) ; 𝑻(𝑛−2) =ℱ𝑥𝑥 (χ𝑛−2 )(𝑝𝑛𝑒𝑡 + 𝑻(𝑛−3) ). (23)

Stage n:
(𝑛) (𝑛−1)
𝐼𝑆𝐼𝑃 (𝑛) = 𝐼𝑆𝐼𝑃′ + 𝑻(𝑛−1) ; 𝑻(𝑛−1) =ℱ𝑥𝑥 (χ𝑛−1 )(𝑝𝑛𝑒𝑡 + 𝑻(𝑛−2) ). (24)

The aforementioned solution can be extended to a multi-cluster fracture design with variable stage spacing
by utilizing the modified influence function ℱ𝑥𝑥 (χ̃𝑖 ) where
2𝑛𝑐 ℎ𝑖
𝜒̃𝑖 = . (25)
𝑛𝑐 +1 (𝐿𝑠 𝑖 +𝐿𝑠 𝑖+1 )

The following is a step-by-step procedure for calculating the hydraulic fracture height for each stage using
the instantaneous shut-in pressure (ISIP) data

1. Calculate the UCS values along the horizontal wellbore. This can be done using seismic AVO
inversion or geomechanical properties derived from drilling data.
URTeC 3860563 11

2. Use the effective instantaneous shut-in pressure (𝐼𝑆𝐼𝑃𝑒𝑓𝑓 ) and minimum horizontal stress derived
from DFIT to calculate the fracture net pressure using 𝑝𝑛𝑒𝑡 = 𝐼𝑆𝐼𝑃𝑒𝑓𝑓 − 𝜎𝑚𝑖𝑛 . The resulting value
of 𝑝𝑛𝑒𝑡 represents the fracture net pressure fracture at the specific point where the DFIT was

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
measured.

3. Establish a linear correlation between the pnet and UCS values for the same Measured Depth (MD),
where the DFIT was conducted. Utilize the derived correlation to estimate the pnet values over the
entire lateral section.

4. Use the drilling derived properties to calculate the minimum horizontal stress along the lateral using
the general Hook’s law. In the event that lateral heterogeneities are insignificant compared to the
vertical, the stress profile along the lateral can be built using the horizontal well tops and the 1-D
vertical stress.

5. Use the calculated pnet and minimum horizontal stress to calculate the virgin ISIP’ along the lateral.
The shut-in pressure (ISIP) obtained from the treatment data for the first stage should match the
calculated virgin shut-in pressure using this method.

6. Calculate the influence function ℱ𝑥𝑥 (𝜒𝑖 ) for each stage (Eq. 1-3) using the actual stage-by-stage
shut-in pressure (ISIP), calculated virgin shut-in pressure (ISIP’) from step 5, and calculated pnet
from step 3.

7. Use the Sneddon equation (Eq. 1) to calculate the value of χi which can be used in ℎ𝑖 = 𝜒𝑖 × (𝐿𝑠 𝑖 +
𝐿𝑠 𝑖+1 ) to estimate the fracture height for each stage.

Note that in the literature, various methods have been proposed to estimate stage-by-stage ISIP from post-
fracture fall-off data. While fitting a straight line to the stabilized portion of the pressure fall-off plot after
the pumps are shut-in is a common approach, more comprehensive methods have been developed to obtain
more accurate estimates of ISIP (e.g., Roussel, 2021). It is important to recognize that the accuracy of the
estimated hydraulic fracture geometry through the proposed method in this paper is heavily reliant on the
quality of the stage-by-stage ISIP measurements. Therefore, it is essential to adopt a reasonable approach
to estimate stage-by-stage ISIP, while carefully considering the wellbore and well completions’ conditions
to minimize the uncertainty associated with ISIP measurements.

2. Quantification of the fracture half-length

2.1 Global fluid continuity

The global fluid volume balance in the fracture is governed by:

𝑉𝑖𝑛𝑗 = 𝑉𝑓𝑟𝑎𝑐𝑡𝑢𝑟𝑒 + 𝑉𝑙𝑒𝑎𝑘 , (26)

where 𝑉𝑖𝑛𝑗 = 𝑄0 𝑡 is the volume of the fluid injected into fracture, 𝑉𝑓𝑟𝑎𝑐𝑡𝑢𝑟𝑒 is the volume of the fluid stored
in the fracture, and 𝑉𝑙𝑒𝑎𝑘 is the cumulative volume of fluid that has leaked off from the fracture.

The cumulative leak-off volume, 𝑉𝑙𝑒𝑎𝑘 , is defined as an integral function of the instantaneous leak-off rate
from the fracture over time and space:
URTeC 3860563 12

𝑡 𝑥𝑓
𝑉𝑙𝑒𝑎𝑘 = ℎ ∫0 ∫−𝑥𝑓 𝑔(𝑥, 𝑡)𝑑𝑥𝑑𝑡 . (27)

The instantaneous leak-off rate can be obtained from 1-D fluid diffusion and is given by (Carter, 1952)

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
𝐶
𝑔(𝑥, 𝑡) =
√𝑡−𝑡0 (𝑥)
. (28)

Here, C is the leak-off coefficient, which can be expressed for a pressure-dependent leak-off as:

2(𝑝𝑓 −𝑝0 )√𝑆√𝑘

𝐶=
√𝜋√𝜇
, (29)

where 𝑝0 is the reservoir pressure, k is the matrix permeability, 𝜇 is the fluid viscosity, and 𝑆 = 𝜙/𝐾 is the
storage coefficient expressed in terms of porosity 𝜙 and bulk modulus K.

The fracture volume, 𝑉𝑓𝑟𝑎𝑐𝑡𝑢𝑟𝑒 can be expressed as an integral of the height-averaged crack opening over
the fracture length:
𝑥
𝑉𝑓𝑟𝑎𝑐𝑡𝑢𝑟𝑒 = ℎ ∫−𝑥𝑓 𝑤
̅(𝑥)𝑑𝑥 , (30)
𝑓

where height-averaged crack opening given by

𝜋ℎ(𝑝𝑓 −𝜎𝑚𝑖𝑛 )
𝑤
̅(𝑥) = . (31)
2𝐸 ′

Here, E is the Young's modulus, 𝜐 is Poisson's ratio, and E' is the effective Young's modulus, which is
expressed as a function of E and 𝜐.

2.2 Local fluid continuity

The flow of fluid inside a PKN fracture is governed by the height-averaged form of the local mass
conservation equation as described by:

𝜕𝑤 𝜕𝑞̅ ̅ 3 𝜕𝑝
𝑤
+𝑔=− ; 𝑞̅ = − , (𝑡 > 0, |𝑥| < 𝑥𝑓 ) (32)
𝜕𝑡 𝜕𝑥 𝜋2 𝜇 𝜕𝑥

Initial and boundary conditions for the constant rate of injection problem are:
𝑄0
𝑝|𝑡=0 = 0, 𝑞̅ |𝑥=0+ = , 𝑞̅ |𝑥=𝑥𝑓 = 0 (33)
2

where the last boundary condition prescribes a no flow boundary condition at the fracture front.

2.3 The solution of the limiting cases of a PKN fracture

The propagation of a hydraulic fracture is influenced by two competing fluid storage mechanisms: the leak-
off dominated regime, where the injected fluid primarily leaks into the surrounding reservoir rocks, and the
storage dominated regime, where most of the injected fluid is stored within the fracture. Additionally, it is
assumed that the dominant dissipative mechanism for a PKN hydraulic fracturing growth is viscosity-
dominated, where viscous energy dissipation prevails over energy dissipation due to rock mass breakage.
This assumption is valid for PKN fractures with lengths much greater than their heights.
URTeC 3860563 13

2.3.1 Limit of leak-off dominated regime (negligible fluid storage)

In the limit of the leak-off dominated regime (𝑉𝑙𝑒𝑎𝑘 ≫ 𝑉fracture), the global mass balance equation can be
simplified as follows

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
𝑉𝑖𝑛𝑗 = 𝑉𝑙𝑒𝑎𝑘 . (34)

Substitution of (28) and (27) in (26) yield:

𝑄0 𝑡 = 𝑉𝑙𝑒𝑎𝑘 = (4ℎ𝐶𝑥𝑓 )√𝑡. (35)

The fracture half-length can be determined by rearranging of (35) expressed as

𝑄0
𝑥𝑓 = √𝑡. (36)
4ℎ𝐶

Shapiro (2006) derived an analytical equation to estimate the size of the triggering front of microseismicity
induced by hydraulic fracturing, given by (Shapiro, 2006)

𝑥𝑓 = √4𝜋𝐷𝑡 (37)

where D is referred to as the fracture diffusivity coefficient. In view of Eq. 36 and Eq. 37, the fracture
diffusivity coefficient can be expressed as (Shapiro, 2006)

𝑄 2
𝐷 = 64𝜋ℎ0 2 𝐶 2 (38)

2.3.2 Limit of the storage dominated regime (negligible fluid leak-off)

The local continuity equation and initial/boundary conditions can be used to quantify the evolution of
fracture half-length in the storage-dominated regime 𝑉fracture ≫ 𝑉𝑙𝑒𝑎𝑘 . Nordgren (1972) derived a solution
for the local continuity equation using the separation of variables technique and given the initial and
boundary conditions.

For a fracture that propagates in the storage-dominated regime, the fracture half-length can be calculated
using the following equation:
1
𝐸’ 𝑄3 5
𝑥𝑓 = 0.34 ( 40 ) 𝑡 4/5 (39)
𝜇ℎ

To extend this solution to a multi-cluster fracture design, we propose a correction factor to be multiplied
by the single cluster fracture half-length. The corrected equation is:
1
𝑛𝑐 +1 𝐸’ 𝑄3 5
𝑥𝑓 = 0.34 ( ) ( 40 ) 𝑡 4/5 (40)
2𝑛𝑐 𝜇ℎ

This correction factor is based on the same analogy as in the case of conversion between a single and multi-
cluster fracture height, which is driven from stress shadowing concept (see section 1.2.1.2). As illustrated
later, the modified xf can generate estimates that are consistent with fracture diagnostic. The maximum
fracture aperture at the fracture inlet can also be determined using the following equation:
URTeC 3860563 14

1
𝑛𝑐 +1 𝜇𝑄02 5 1/5
𝑤
̅(x = 0) = 2.5 ( )( ) 𝑡 (41)
2𝑛𝑐 𝐸’ ℎ

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
2.3.3 Importance of the leak-off dominated regime in the hydraulic stimulation of tight reservoirs

This section investigates the importance of the leak-off-dominated regime relative to the storage-dominated
regime during a typical hydraulic fracturing operation in the Montney Formation. To estimate the 1-D leak-
off coefficient, typical reservoir and fluid parameters are assumed, including a reservoir permeability of
𝑘 = 100 nd, fracture pressure 𝑝𝑓 = 50 MPa, 𝑝0 = 28 MPa, reservoir storage parameter S = 5 × 10−11
1/Pa, and fluid viscosity 𝜇 = 1 cp. Based on these parameters, the 1D pressure-dependent leak-off
coefficient is calculated to be 𝐶 = 1.8 × 10−6 m/√𝑠 (Eq. 29).

Next, assuming a hydraulic fracture height of h=60 m, hydraulic fracture half-length of xf = 400 m, a 10-
cluster fracture stage design, and a total injected volume per stage of 𝑉𝑖𝑛𝑗 = 800 m3, the fluid leak-off
volume is calculated from Eq. 27 to be at 𝑉𝑙𝑒𝑎𝑘 =115 m3.

The fluid efficiency (FE), defined as the percentage of fluid stored in the fracture divided by total injected
fluid, is then calculated as 𝐹𝐸 = (𝑉𝑖𝑛𝑗 − 𝑉𝑙𝑒𝑎𝑘 )/𝑉𝑖𝑛𝑗 , resulting in an FE of approximately 85%. This
calculated fluid efficiency is consistent with typical fluid efficiencies estimated from DFITs in the Montney
Formation.

It should be noted that Kovalyshen and Detournay (2010) proposed a more rigorous approach to calculate
a characteristic timescale associated with the fluid leak-off process given by
2/3
𝜋3 𝜇𝑄02
𝑡𝑚𝑚̃ = ( ) (42)
2𝐸 ’ 𝐶 5 ℎ

where m and m
̃ are associated with storage and leak-off dominated regimes.

In view of Eq. 42, the leak-off regime is dominant when the injection time exceeds the characteristic time
scale, denoted as 𝑡 ≫ 𝑡𝑚𝑚 ̃ . Based on calculations, the characteristic time scale for a typical hydraulic
fracturing job is estimated to be approximately 𝑡𝑚𝑚 ̃ = 287 days. This suggests that the injection time is
much shorter than the characteristic time scale needed for the leak-off regime to take over the storage regime
during hydraulic stimulation of a tight reservoir. Therefore, it is reasonable to use the solution for the
limiting case of a fracture propagating in the storage-dominated regime to estimate the fracture half-length
and width in unconventional tight reservoirs.

3. Case Study: Crosscutting well in Middle Montney Formation

A 6-well pad development program was undertaken in the Montney Formation of Northeast British
Columbia, which included the drilling of a crosscut horizontal well, along with two horizontal wells in the
Upper Montney, two in the Middle Montney (one of which is the crosscut well), and two in the Lower
Montney, with a lateral spacing of 400 meters between wells in the same bench. A microseismic program
was conducted to monitor the hydraulic fracturing geometry of the pad, using geophones placed near the
heel of a Lower Montney well in the proximity of the pad of interest to ensure sufficient coverage of the
microseismicity was established as depicted in the Figure 8.
URTeC 3860563 15

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Figure 8: Schematic of pad of interest in the Montney. The pad was monitored using microseismic from the offset parent pad.

3.1 Geology and geomechanics in well of interest

The Lower Triassic Montney Formation is composed of three primary stratigraphic sequences that are
bounded by either boundary–flooding surfaces or their correlative conformities. In order, these include
sequence 1 (Lower Montney), sequence 2 (Middle Montney), and sequence 3 (Upper Montney). Based on
work completed by Baniak et al. (2023), sequence 2 within northeastern British Columbia can be
subdivided into four parasequence sets (MmA–MmD) and one separate Upper parasequence (MmE).
Notably, MmC and MmD parasequence sets are mostly composed of a mixed or hybrid lithology of
interbedded siliciclastic and carbonate bioclastic deposits. Sequence 3 is composed of five parasequence
sets (Um1 to UmC, as per Baniak et al. 2023) and is notably devoid of any carbonates.

The lithology and rock composition of the parasequence sets in this crosscut well exhibits unique variations,
leading to heterogeneous geomechanical properties. MmC and MmD parasequence sets, for instance,
exhibit high mechanical stiffness. Much of this is likely due to the presence of calcite-cemented bioclasts,
resulting in an exceptionally hard rock. DFIT measurements also confirm that these Middle Montney
carbonates packages are significantly overpressured and overstressed. Conversely, the Um1 parasequence
set appears to be mechanically hard due to the abundance of quartz minerals and absence of carbonate,
leading to higher porosity and improved brittleness.

The crosscut horizontal well example presented in Figure 9, it was drilled starting from the Middle Montney
MmC parasequence set with the heel section located at a depth of approximately 2000 m TVD. The well
was drilled toe-up, cutting through several stratigraphic units before reaching the bottom section of the Um1
parasequence set.
URTeC 3860563 16

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Figure 9: Cross sectional view of the crosscut well in the Montney formation

A comprehensive rock physics study was conducted to establish a robust correlation between the
mechanical properties, pore pressure, and in-situ stresses obtained from various data sources, including log
measurements, direct core test studies, and DFIT measurements. Using these data, a high-fidelity 1-
Dimensional (1D) Mechanical Earth Model (MEM) was constructed for the area of interest, as depicted in
Figure 10, providing critical insights into the subsurface geomechanics of the area. The red box in the figure
highlights the vertical depth of the crosscut well overlaid on the 1D MEM model.

The unique set of geomechanical properties exhibited by the Middle Montney sequence (MmC, MmD,
MmM) results in highly uncontained hydraulic fractures with a large fracture height. As a result, completing
wells in the Middle Montney presents significant challenges, including difficulty in placing proppant, a
phenomenon known as tough fracturing, as well as potential casing deformation. Additionally, post-
fracturing image logging of the Middle wells often reveals poor cluster efficiency due to potential stress
escalation, also known as stress shadowing, which can result in competing fractures within a stage. In
contrast, the quartz dominated Um1 sequence exhibits smaller fracture height and is typically associated
with fewer operational fracturing challenges. The microseismic campaign was designed to provide insights
on the variability of the hydraulic fracturing geometry drilled through multiple notoriously heterogeneous
geomechanical zones.
URTeC 3860563 17

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Figure 10: 1D Mechanical Earth Model (MEN) of well of interest constructed using various data sources, including log measurements, direct core
test studies, and DFIT measurements.

3.2 Estimating stage-by-stage fracture height and length in the well of interest

In this section, we utilize the methodology described in Section 1.2.2 to quantify the variability in
the hydraulic fracture height along the lateral. Figure 11 illustrates the variation of the ISIP across
different stages in the crosscut well. The ISIP value demonstrates a clear upward trend with
increasing stage number, as shown in the Figure. However, the variability in geomechanical
properties has resulted in non-uniform hydraulic fracture height, causing the ISIP curve to deviate
from the expected exponential shape that assumes uniform fracture height and geomechanical
properties.

Figure 11: ISIP as function of stage number in the crosscut well.

URTeC 3860563 18

In the first step of the analysis, we utilized an internal algorithm to calculate the UCS values along
the horizontal wellbore. This algorithm involved integrating and analyzing drilling parameters, such
as penetration rate (ROP), torque, and bit type. An empirical correlation was used to relate these

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
parameters to rock strength, elastic properties, in-situ stress, and reservoir pore pressure, enabling
estimation of these geomechanical properties from drilling data.

Figure 12 shows the variation of UCS, and pore pressure (pp) estimates along the lateral, represented
by the black line, using drilling-derived data. The red line represents the same parameters derived
from the horizontal well tops and the 1-D MEM model.

Figure 12: Variation of Unconfined Compressive Strength (bottom figure) and pore pressure (top figure) estimates along the lateral
using drilling-derived data. The red line represents the same parameters derived from the horizontal well tops and the 1D MEM model.

In the subsequent step, a linear correlation was established between the 𝑝𝑛𝑒𝑡 and UCS by utilizing a
single DFIT measurement as a calibration point. The virgin shut-in pressure (ISIP’) was then
calculated for the entire lateral length, using the 𝑝𝑛𝑒𝑡 and minimum in-situ stress obtained along the
horizontal borehole. Figure 13 presents a visual representation of the distribution of ISIP’ and
𝑝𝑛𝑒𝑡 . The values of the minimum horizontal stress along the lateral well are illustrated in Figure 14.

Figure 13: The virgin ISIP’ and 𝑝𝑛𝑒𝑡 as function of the stage number in the crosscut well
URTeC 3860563 19

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Figure 14: Minimum horizontal stress as function of the stage number in the crosscut well

In the final step, the influence function, denoted as ℱ𝑥𝑥 (𝜒𝑖 ), can be calculated for each stage using
the actual stage-by-stage shut-in pressure, the calculated virgin shut-in pressure (ISIP’), and the
calculated net pressure (pnet). Once the influence function is obtained, the value of χ𝑖 can be
calculated and used in the equation ℎ𝑖 = 𝜒𝑖 × (𝐿𝑠 𝑖 + 𝐿𝑠 𝑖+1 ) to estimate the fracture height for each
stage, as illustrated in Figure 15. The estimated values of the fracture height from microseismic are
overlaid on the same graph for comparison. Note that each stage was ~100 m long on average.

Figure 15: Comparison of fracture height estimates obtained through ISIP analysis (in red) and microseismic-derived height values (in
blue)

After determining the fracture height, the average fracture half-length can be calculated using Eq.
40. It was noted that microseismic results in the lateral direction were biased towards the location of
the geophones (i.e., towards the SW). Figure 16 provides a comparison between the ISIP results and
the average hydraulic fracturing length estimated from microseismic data (averaged between NE
and SW sections). Moreover, Figure 17 demonstrates that the ISIP results align more closely with
the half-length of fractures in the SW section.
URTeC 3860563 20

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Figure 16: Comparison of fracture half-length estimates obtained from ISIP analysis (in red) and the averaged half-length values
derived from microseismic data between NE and SW sections (in blue)

Figure 17: Comparison of fracture half-length estimates obtained from ISIP analysis (in red) and microseismic-derived averaged half-
length in SW direction (in blue)

The comparison between the seismic deformation and calculated fracture half-length from ISIP is
also presented on the map view in Figure 18.

The crosscut has an estimated median fracture height of ℎ ≈ 110 m and half-length 𝑥𝑓 ≈ 380 m
estimated from ISIPs. To compare the average hydraulic fracture height, Roussel’s work is used,
assuming negligible geomechanical heterogeneity along the crosscut. Although it may not be clear
from the ISIP curve where the stress plateau occurs, it is assumed to be around 𝛥𝜎𝑥𝑥 Plateau ≈ 6.5
MPa. With an average fracture net pressure of pnet ≈ 3 MPa (averaged over the entire lateral based
on the approximation in Figure 13, the estimated fracture height is approximately around h ≈130 m.
Using the analytical solution for simplicity (Eq. 9), the estimated fracture height is approximately
h ≈160 m.
URTeC 3860563 21

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Figure 18: Comparison of seismic deformation (in Green) with fracture half-length calculated through ISIP analysis (in Red).

As demonstrated in this section, the proposed approach allows for rapid estimation of the variability
of the average hydraulic fracturing length and height, which is found to be reasonably consistent
with results obtained from microseismic monitoring.

3.3 Fracture modeling in the well of interest

A hydraulic fracture model was developed using Kinetix Stimulation Software Suite, incorporating
geological and reservoir data, fracture properties, and simulation parameters. The model results
show that fractures in the MmD and MmC have a strong upward growth tendency with limited
fracture length and can breach the MmE sequence to propagate into the Upper Montney and
occasionally into the Doig Formation as illustrated in Figure 19. Moreover, propagating fractures in
higher stressed zones led to extreme events in fracture height growth and stronger stress shadowing
effects that impacted cluster efficiency and far-field fracture uniformity. The lack of fracture
containment can be undesirable limiting lateral fracture extension and hinder proppant transport,
reducing productivity. However, the containment of fractures was found to improve in MmC
compared to MmD, as expected due to the stress pinch in the MmD and the transition to lower stress
zones closer to the bottom of the MmC.

Figure 19: Schematic of vertical fracture growth in the crosscut well using fracture modeling.
URTeC 3860563 22

The hydraulic fracturing modeling results also indicate that fractures in the Upper Montney Um1
have smaller fracture heights, but larger fracture half-lengths compared to those in the Middle
Montney, and remain contained within the Upper Montney. The MmD parasequence is typically

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
effective at preventing fractures from the Upper Montney zone from propagating downwards. These
behaviors in hydraulic fractures were confirmed through additional methods such as the Volume-
To-First Response (VFR) in the Sealed Wellbore Pressure Monitoring (SWPM) and Interference
Testing in the Montney Formation (Chin et al.).

The map view in Figure 20 shows a comparison between the seismic deformation and the calculated
fracture length using fracture modeling. The consistency of results across multiple datasets and
fracture diagnostic methods is crucial for accurately characterizing the reservoir and optimizing
production in unconventional Montney rocks.

Figure 20: Comparison of seismic deformation with fracture half-length obtained through fracture modeling.

Conclusions
Optimizing the productivity and efficiency of unconventional rocks requires a thorough
understanding of hydraulic fracture geometry. However, developing this understanding can be
challenging due to several factors, including the inherently heterogeneous rock masses with complex
stress regimes, as well as competing fractures in multi-cluster, multi-stage fracture jobs created from
different benches of a multi-well pad. This work aimed to develop a fast and robust framework for
quantifying the variability of hydraulic fracturing height and length during multi-cluster, multi-stage
hydraulic fracturing treatments. To achieve this, a framework was proposed to utilize the stage-by-
stage ISIP analysis, accounting for lateral geomechanical heterogeneity along the horizontal
borehole. The proposed methodology was applied to a horizontal well drilled through multiple
notoriously heterogeneous geomechanical zones in the Montney Formation, monitored by a
downhole microseismic program. It was demonstrated that this computationally fast approach can
replicate the variability of the average hydraulic fracturing length and height at relatively no cost,
while being in good agreement with microseismic monitoring. This approach can be used alone or
in combination with other fracture diagnostic tools to help calibrate fracture modeling and provide
crucial insights for optimizing production in unconventional rocks.
URTeC 3860563 23

Acknowledgements
The author gratefully acknowledges PETRONAS Energy Canada Ltd. for granting permission to
publish this article. The author also thanks colleagues at PETRONAS Energy Canada Ltd. for

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
support, encouragement, and valuable insights and feedback on the article.
References
1. Baniak, G.M., T.F. Moslow, S. Michailides, and M.G. Adams. Sequence stratigraphic
architecture of the Lower Triassic Montney Formation, northeastern British Columbia,
Canada. AAPG Bulletin 107, no. 2 (2023): 283-310.
2. Carter, E., 1957. Optimum fluid characteristics for fracture extension, in: Howard, G.C., Fast,
C.R. (Eds.), Drilling and Production Practices, pp. 261–270.
3. Chin, Adam, P. Miller, D. Redpath, K. Dauncey, D. Nakaska, and F. Alimahomed. Sealed
Wellbore Pressure Monitoring in the Montney: Early Learnings. In SPE Canadian Energy
Technology Conference. OnePetro, 2022.
4. Cipolla, C.L., M.J Williams, X. Weng, M. Mack, and S. Maxwell. Hydraulic fracture monitoring
to reservoir simulation: maximizing value. In Second EAGE middle east tight gas reservoirs
workshop, pp. cp-244. EAGE Publications BV, 2010.
5. Cipolla, C. L., and C. A. Wright. Diagnostic techniques to understand hydraulic fracturing:
what? why? and how? In SPE/CERI Gas Technology Symposium. OnePetro, 2000.
6. Economides, M.J., and K.G. Nolte. Reservoir stimulation. Vol. 2. Englewood Cliffs, NJ:
Prentice Hall, 1989.
7. Garagash, D. I. Cohesive-zone effects in hydraulic fracture propagation. Journal of the
Mechanics and Physics of Solids, 133, 103727. (2019)
8. Haustveit, K., B. Elliott, and J. Roberts, (2022, January). Empirical Meets Analytical-Novel
Case Study Quantifies Fracture Stress Shadowing and Net Pressure Using Bottom Hole Pressure
and Optical Fiber. In SPE Hydraulic Fracturing Technology Conference and Exhibition.
OnePetro.
9. Kovalyshen, Y., and E. Detournay. A reexamination of the classical PKN model of hydraulic
fracture. Transport in porous media 81 (2010): 317-339.
10. Mayerhofer M.J., E.P. Lolon, N.R. Warpinski, C.L. Cipolla, D. Walser, and C.M. Rightmire.
What is stimulated reservoir volume? SPE Production & Operations 25, no. 01 (2010): 89-98.
11. Montgomery, C.T., and M.P. Smith. Hydraulic fracturing: History of an enduring technology.
Journal of Petroleum Technology, vol. 66, no. 5, 2014, pp. 28-40.
12. Moradi, P. M. (2021). SERA: An ISIP analysis approach to estimate fracture height and
correlate stress escalation and relaxation in multifrac horizontal wells. Results in
Engineering, 11, 100241.
13. Nolte, K. G., 1991, Fracturing-Pressure Analysis for Nonideal Behavior, J. Pet. Technol., 43(2),
pp. 210–218.
14. Nordgren, R. P. (1972). Propagation of a vertical hydraulic fracture. Society of petroleum
engineers journal, 12(04), 306-314.
15. Roussel N.P. Analyzing ISIP stage-by-stage escalation to determine fracture height and
horizontal-stress anisotropy. In SPE Hydraulic Fracturing Technology Conference and
Exhibition. OnePetro, 2017.
URTeC 3860563 24

16. Roussel, N., Swan, H., Snyder, J., Nguyen, D., Cramer, D. and Ouk, A., 2021, December.
Evaluation and Insights from Instantaneous Shut-in Pressures. In Unconventional Resources
Technology Conference, 26–28 July 2021 (pp. 2545-2563). Unconventional Resources

Downloaded from http://onepetro.org/URTECONF/proceedings-pdf/23URTC/3-23URTC/D031S071R003/3182738/urtec-3860563-ms.pdf by King Fahd University of Petroleum & Minerals user on 24 December 2024
Technology Conference (URTeC).
17. Sarvaramini, E., and D.I. Garagash. Breakdown of a pressurized fingerlike crack in a permeable
solid. Journal of Applied Mechanics 82, no. 6 (2015).
18. Shapiro S.A., C. Dinske, and E. Rothert. Hydraulic‐fracturing controlled dynamics of
microseismic clouds.Geophysical Research Letters 33.14 (2006).
19. Sneddon, I. N. (1946). The distribution of stress in the neighbourhood of a crack in an elastic
solid. Proceedings of the Royal Society of London. Series A. Mathematical and Physical
Sciences, 187(1009), 229-260.
20. Warpinski, N.R., M.J. Mayerhofer, M.C. Vincent, C.L. Cipolla, and E.P. Lolon, Stimulating
unconventional reservoirs: Maximizing network growth while optimizing fracture conductivity.
Journal of Petroleum Technology, vol. 60, no. 12, 2008, pp. 84-93