DC191335 DOI: 10.

2118/191335-PA Date: 31-January-19 Stage: Page: 1 Total Pages: 11

Evaluating Cement-Plug Mechanical

and Hydraulic Integrity
Axel-Pierre Bois, Manh-Huyen Vu, Kim Noël, and Anthony Badalamenti, CURISTEC;
Laurent Delabroy, AkerBP; and Emmanuel Thérond and Knut Hansen, BP

Summary
In past years, the industry has focused on ensuring that cement is efficiently placed in the wellbore and that it does not become mechan-
ically damaged during the life of the well. However, few efforts have been made to determine how cement mechanical integrity (CMI)
relates to cement hydraulic integrity (CHI) (i.e., evaluating the flow rate that could occur through the cement barrier), even though CHI
is one of the main objectives of placing a cement plug in a wellbore.
The analysis of hydraulic integrity requires that a CMI model be used to compute the state of stress and pore pressure in the cement
and to estimate which type of mechanical failure might occur during the life of the well. It also requires that a CHI model be integrated
with the CMI model to estimate the rate of fluid that might flow through a cement barrier, should it mechanically fail. This provides the
engineer with insight into the long-term integrity of a cement plug.
This paper describes the work conducted on CMI/CHI models for cement plugs, and it presents a sensitivity analysis that demon-
strates the value of an integrated CMI/CHI model. The study indicates that (1) well geometry, cement properties, reservoir pressures,
cement heat of hydration, and fluid properties are required inputs for proper analysis; (2) the changes of stresses and pore pressure over
time need to be computed along the length of the cement plug, with sensitivity analysis to consider the existing uncertainties; (3) a
cement plug might preserve its sealing capability, even if the CMI model shows the existence of a microannulus (e.g., when the fluid
viscosity is very high); and (4) a cement plug might lose its sealing capacity, even if the CMI model shows no induced defect
(e.g., when a microannulus is propagated as a hydraulic fracture).
These last two observations are important because they show that what a CMI model cannot predict, a CHI model can.

Introduction
Well abandonment is the final stage in the life of a well. In many countries, it is mandatory to re-establish the natural integrity of the
formations that have been crossed by the well. If the well is not correctly abandoned, fluid might flow through the cement plug, which
could lead to fluid migration between reservoirs or to the surface.
Well abandonment is generally performed by placing cement plugs across natural geological seals to restore the isolation of the res-
ervoirs. Different methods of plug placement for well abandonment can be used, such as the perforate, wash, and cement (PWC) tech-
nology (Ferg et al. 2011; Delabroy et al. 2017). Daccord et al. (2006) presented an extensive discussion of the state-of-the-art on this
topic. In this paper, it is assumed that the cement plugs are placed as per design.
Our investigation revealed that there is very limited published literature on the mechanical design of cement plugs with the objective
of preventing fluid-flow paths for the life of the cement plug. Akgun and Daemen (1994, 1997) performed mechanical and hydraulic-
integrity experiments with cement plugs placed in steel and basalt cylinders. Nagelhout et al. (2010) presented laboratory experiments
on the performance of two cement slurries, showing that the resilient, expansive cement system performs better than the resilient, non-
expansive cement system. Opedal et al. (2018) developed an experiment to simulate cement-plug CHI in the laboratory. The baseline
experiments were defined as dry experiments with American Petroleum Institute (API) Class G cement without additives. The sealing
efficiency of the plug was tested with gas. These experiments showed that placing a water column on top of the cement column gives a
fluid-flow breakthrough at a higher pressure compared with when no such water column is present. However, none of these works pre-
sented a design methodology. Mainguy et al. (2007) performed numerical simulations of the consequences of reservoir production on
the stability of cement plugs, showing that the main risk of a conventional API Class G cement system comes from tensile cracks,
whereas resilient and flexible cement systems perform better to prevent cement damage. However, cement hydration was not simulated,
and no design methods were presented.
The standard industry practice typically consists of evaluating the integrity of the cement plug after it has set by performing a posi-
tive or a negative pressure test. The positive pressure test consists of applying a pressure to the top of the plug and checking if there is a
decrease in applied wellhead pressure (WHP) after the wellhead is closed. This test indicates the sealing efficiency of the plug against a
downward fluid flow and not against an upward fluid flow, which could occur if the cement plug encounters pressure from reservoirs
below. In this latter case, one should perform a negative pressure test, which consists of decreasing the pressure above the plug and
checking if an upward fluid flow could occur.
The objective of this paper is to present a model dedicated to simulating the mechanical and hydraulic integrity of a cement plug,
after the cement has set, to understand what can be done to prevent cement-plug damage in the short and long term.

Cement-Hydration Model
The cement clinker is composed of four main phases: calcium silicates (C3S, C2S), calcium aluminates (C3A), and calcium aluminofer-
rites (C4AF), where, in the standard cement chemistry, the notation C stands for CaO, S for SiO2, A for Al2O3, and F for Fe2O3. The
clinker is pulverized together with calcium sulfate, usually gypsum, to form Portland cement. In contact with water, these anhydrous
components dissolve, and the resulting ions precipitate in the form of calcium silicates hydrates (C-S-H) and Portlandite [Ca(OH)2]. At
a temperature higher than 110 C (230 F), the C-S-H phase often converts to a phase called alpha dicalcium silicate hydrate (a-C2SH),
leading to a decrease in cement compressive strength, a phenomenon called strength retrogression (Nelson and Barlet-Gouédard 2006).

Copyright V
C 2019 Society of Petroleum Engineers

This paper (SPE 191335) was accepted for presentation at the SPE Norway One Day Seminar, Bergen, Norway, 18 April 2018, and revised for publication. Original manuscript received for
review 13 March 2018. Revised manuscript received for review 9 July 2018. Paper peer approved 16 July 2018.

2019 SPE Drilling & Completion 1

ID: jaganm Time: 15:42 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 2 Total Pages: 11

To prevent this, the bulk lime/silica ratio in the cement is generally reduced by partially replacing some Portland cement by ground
quartz, usually fine silica sand or silica flour (Nelson and Barlet-Gouédard 2006).
Cement hydration is conducted by a series of complex exothermic, thermo-activated chemical reactions, which are associated, to
some degree, with volumetric shrinkage, such that the volume of the products is smaller than the volume of the reactants (Powers and
Brownyard 1947). This gives rise to intergranular porosity, which also exists in the C-S-H grains. Therefore, set cement is a porous
material and should be simulated as such (Bois et al. 2011).
However, most CMI simulators do not explicitly model cement hydration to compute the state of stress and pore pressure in the
cement after it has set. They are based on a continuous, one-phase, linear-elasticity theory up to a threshold that depends on the
cement’s unconfined compressive and tensile strengths. They do not consider the set-cement thermo-poro-mechanical behavior. This
means that most CMI simulators do not follow the first three requirements for an effective CMI simulator as described by Bois et al.
(2015)—that is, the capacity to (1) evaluate the initial state of stress in the cement after it has set; (2) estimate where the cement design
lies in the safe-working envelope and how close it is to the envelope limits; and (3) simulate the stiffness and failure behavior of all
materials to correctly locate the failure-criterion limits. Consequently, these simulators cannot capture all cement-integrity features.
After cement is placed in the wellbore, its hydration degree and properties evolve with time, resulting from various chemical reac-
tions. This evolution depends on temperature, stresses, and water exchanges between the cement and the formation. A hydration model
that allows simulating this hydration process has been developed and validated. It allows computing various cement properties as a
function of the degree of hydration or as a function of time, such as porosity, shrinkage, elastic properties (Young’s modulus, Poisson’s
ratio), plastic behavior, failure properties (cohesion, friction angle), and heat-transfer properties (conductivity, specific heat).
The model presented here is based on the following hypotheses (the definition of the symbols is given in the Nomenclature section):
• Cement hydration is monitored through a hydration degree (n) that is defined as the ratio of produced heat to total heat produced
when all cement is hydrated.
• Cement thermo-activated hydration is modeled on the basis of a macroscopic kinetic law:
 
dn ~  exp Ea : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð1Þ
¼ AðnÞ
dt RT
• The chemical affinity depends not only on cement composition and Blaine fineness of cement powder (ASTM C204-18 2018), but
also on the diffusion of water through hydrate layers and the actual hydration degree. The effects of temperature and pressure
are considered.
• Cement exothermic hydration is modeled on the basis of the following heat-transfer equation:
dT dn
C ¼ divðk  gradTÞ þ q1  : . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð2Þ
dt dt
• The hydration degree evolves vs. time, and it stops at a maximum value that is a function of the water/cement (w/c) ratio
and temperature.
• During hydration, hydrates are formed around anhydrate grains to form larger grains. At the beginning, the slurry behaves as a
suspension of grains in water. At a given porosity, these grains start to contact with adjacent ones. The related degree of hydration
is called the percolation threshold. Its value depends on cement composition, w/c ratio, and the type of additives. Inert additives
are simulated through the theory of inclusions (Mura 1987; Dormieux et al. 2006), whereas reactive additives modify the kinetics
of cement hydration.
• Swelling/shrinkage of cement comes from hydration heat (thermal expansion), water consumption by chemical reactions (leading
to a decrease in pore pressure), cement grain-volume increase, and water exchange with the formation.
• Cement behavior after the percolation threshold is simulated according to thermo-chemo-poro-elasto-plasticity.
• Cement thermo-poro-elastic law, when cement is a porous solid saturated with a fluid with density qf, is written as
 
2G
dr ¼ Kd   trde  I þ 2G  de þ b  dpp þ 3Kd  ðsd  dn þ ad  dTÞ  I
3
!
dm Ku  au  Kd  ad Ku  su  Kd  sd
dpp ¼ M  b  trde þ þ3  dT  3  dn:                          ð3Þ
qf b b

• Cement plasticity is simulated in the frame of pore-collapse theories.

Consequently, it is not necessary for the engineer to estimate the initial state of stress in the cement after hydration because this ini-
tial state of stress is directly computed. Moreover, it is possible to evaluate an initial state of stress before reaching a plateau, in case the
engineer would prefer to evaluate the impact that a well operation would have on the cement integrity at a very young age in the well.
Thérond et al. (2017) presented the validation of this CMI model after noting three sets of experimental data. Therefore, it is not
reproduced in this paper.

Cement-Plug Model
A cement-plug CMI model has been developed by solving Eqs.1 through 3 for a cement plug in a hole with the following
configurations:
• The plug is in an open hole or a cased hole.
• The cement is saturated by water or is under partial-saturation conditions because of the water consumption that takes place during
cement hydration.
• The interface between the plug and the hole is perfectly closed, or is open if a microannulus is induced.
• Simulations are performed under plane strain when there is no microannulus and at a distance from the top and the bottom of the
cement plug.
• Axial sliding of the cement plug is considered when a microannulus opens.
Furthermore, a CHI module has been developed in which the flow of fluid in the microannulus is simulated, while considering the
coupling between the hydraulics (fluid pressure) and the mechanics (solid-media stiffness), in the same way as it is conducted for
hydraulic-fracturing operations. Four shear-dependent viscosity models are available, [i.e., Newton, Ostwald-De Waele, Bingham, and
Herschel-Bulkley (Guillot 2006)], and losses of pressure in the microannulus are simulated (Fig. 1).

2 2019 SPE Drilling & Completion

ID: jaganm Time: 15:42 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 3 Total Pages: 11

Microannulus

ΔP

e1
δpf

Cement plug e2 Formation or

casing

Fig. 1—Simulation of pressure loss in a microannulus: (DP) denotes the applied pressure, (dpf) is the shear stress at the faces of
the microannulus, and (e1) or (e2) is the aperture of the microannulus at different locations.

The volume of the wellbore fluid above the plug is considered to simulate pressure tests (i.e., to compute the decrease or increase of
WHP induced by a downward or upward flow).
A sensitivity analysis that is based on this model is presented in the next subsections, which include (1) an analysis of the process of
positive pressure-testing a cement plug; (2) a study of how various parameters affect the ability of a cement plug to maintain its hydrau-
lic integrity; and (3) a four-well comparison between the model prediction and data recorded during pressure tests. Finally, some con-
clusions are given on some methods to help ensure cement-plug integrity.

Positive Pressure Testing a Cement Plug. A positive pressure test is generally performed by increasing the WHP to a value such that
the pressure applied at the top of the plug is greater than the fracture-initiation pressure below the cement-plug barrier by a given
margin, and by checking if the pressure-loss rate during a fixed time period remains below a fixed threshold, which is the maximum-
allowable pressure-loss rate (PLR0). For example, the test could use a margin of 3.45 MPa (500 psi) during a time period of 30 minutes,
and it could check if the pressure loss remains below 34.5 kPa/min (5 psi/min) during this time period.
However, such a test assumes that there is an acceptable decline rate in terms of pressure and not in terms of fluid mass or fluid
volume, which means that the maximum-allowed volume-loss rate (VLR0) depends on the operational conditions.
To illustrate this point, we have computed the equivalent VLR0 as a function of PLR0 for a cement plug in a cemented casing, under
different conditions. The base case is for a 0.2445-m (95=8-in.) casing cemented in a 0.3111-m (121=4-in.) hole. The cement and forma-
tion Young’s-modulus and Poisson’s-ratio values are the same {i.e., 10 GPa [1.45 Mpsi (million psi)] and 0.2}. The mud on top of the
cement has a specific gravity (SG) of 1 g/cm3 (8.35 lbm/gal) and a compressibility of 0.001/MPa (6.9 106/psi). A pressure of 7 MPa
(1,015 psi) is applied at the wellhead, and the PLR0 is 0.035 MPa/min (5 psi/min).
These computations show that the presence of cement behind the casing has only a slight impact on the computation of the equiva-
lent VLR0 for a time period of 1 minute, whereas the depth of the top of the cement, the mud compressibility, the casing size, and the
PLR0 have a significant impact on the equivalent VRL0 (Fig. 2). They also showed that the SG of the mud and cement, and the
Young’s modulus of the formation, have a negligible impact (figures not shown for the sake of brevity).

Cement-Plug Hydraulic Integrity. To illustrate the use of the cement-plug mechanical and hydraulic-integrity simulation tool, we
present here a case study of a cement plug that is set in a vertical openhole section between a 728- and 978-m (2,388- and 3,209-ft)
measured depth.

Inputs to the Case Study. In the base case, the hole has a diameter (HD) of 0.216 m (81=2 in.), and the formation has Young’s-modulus (E)
and Poisson’s-ratio () values of 5 GPa (0.725 Mpsi) and 0.20, respectively. The cement is a 1.95-g/cm3 (16.3-lbm/gal) API Class G
cement system with Young’s-modulus (E) and Poisson’s-ratio () values of 10 GPa (1.450 Mpsi) and 0.14, and a heat-of-hydration (Q)
value of 604 MJ/m3 (16,211 Btu/ft3). The SG (q) of the mud above the plug is 1.7 g/cm3 (14.2 lbm/gal). It follows a Bingham viscosity
model with a 4.79-Pa (10-lbf/100 ft2) yield stress (sy) and a 0.045-Pas (0.094-lbf-sec/100 ft2) consistency index (k). The applied WHP
during the pressure test is 10 MPa (1,450 psi).
Table 1 presents a synthesis of the results of all simulations in this section.

Simulation-Result Conventions. The simulations show that the most important risk for losing cement-plug integrity is through the
creation of a microannulus at the cement-plug/hole interface. The model provides support that other risks, such as damaging the cement
in compression or tension, are much less significant.
The simulation outputs are represented in Figs. 3 through 6 where the horizontal and vertical axes denote the size of the microannu-
lus and the depth. A negative number for the microannulus aperture means that the two faces of the interface between the cement and
the wellbore overlap and a microannulus has not been created, whereas a positive number indicates that a microannulus has been cre-
ated, the value being equal to the opening in mm.

Base Case. Fig. 3a presents the geometry of the microannulus at the interface between the cement plug and the wellbore just after
hydration, on the basis of the CMI simulation, which does not check the risk of propagating a microannulus such as a hydraulic fracture.
The values are always negative, which indicates that a microannulus has not been created.

2019 SPE Drilling & Completion 3

ID: jaganm Time: 15:42 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 4 Total Pages: 11

Influence of the Existence of a Cements Heath on VLR0 Influence of Casing Size on VLR0
30 140
Cemented 4½ in.
25 Uncemented 120 7 in.
9⅝ in.
100
13⅜ in.
VLR0 (liter)

20

VLR0 (liter)
20 in.
80
15
60
10
40

5 20

0 0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Depth of the Top of Cement Plug (m) Depth of the Top of Cement Plug (m)
(a) Existence of cement behind the casing (b) Casing size

Influence of PLR0 on VLR0 Influence of Mud Compressibility on VLR0

60 200
5 kPa 0.0001/MPa
180
50 20 kPa 0.0005/MPa
160
35 kPa 0.001/MPa
40 50 kPa 140 0.005/MPa
VLR0 (liter)

VLR0 (liter)
65 kPa 120 0.01/MPa
30 100
80
20
60
40
10
20
0 0
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Depth of the Top of Cement Plug (m) Depth of the Top of Cement Plug (m)
(c) PLR0 (d) Mud compressibility

Fig. 2—Influence of various parameters on the VLR0.

Inputs Microannulus (μm) WHP change

Hole Formation Cement Mud At Top of Plug At Bottom of Plug (kPa/min) (psi/min)

Size E E Q τy k Phase Phase Phase Phase Phase Phase Phase Phase Phase Phase
(in.) (GPa) (GPa) (MJ/m3) (Pa) (Pa⋅s) 1 2 3 1 2 3 2 3 2 3

8½ 5 10 604 4.8 0.045 –255 216 597 –406 –406 117 0 1.40 0 0.20

8½ 3.75 10 604 4.8 0.045 –359 217 684 –551 –551 138 0 2.35 0 0.34

8½ 6.25 10 604 4.8 0.045 –193 214 544 –320 –320 109 0 0.98 0 0.14

12¼ 5 10 604 4.8 0.045 –367 278 801 –580 –580 –580 0 4.05 0 0.59

8½ 20 10 604 4.8 0.045 –34 199 387 –93 –93 136 0 0.31 0 0.04

8½ 5 10 604 9.6 0.045 –255 216 597 –406 –406 –406 0 0 0 0.00

8½ 5 10 604 4.8 0.090 –255 216 597 –406 –406 117 0 0.69 0 0.10

8½ 5 7 604 4.8 0.045 –240 296 731 –405 –405 161 0 4.82 0 0.70

8½ 5 10 300 4.8 0.045 –249 253 659 –406 –406 –406 0 2.63 0 0.38

E: Young’s modulus; Q: heat of hydration; τy: yield stress; k: consistency index; Phase 1: CMI after hydration; Phase 2: CHI after hydration; Phase 3: CHI during testing.
A negative-aperture microannulus means that the two faces of the interface overlap; a positive-aperture microannulus is equivalent to an opening. In bold are input data
modified from the base case.

Table 1—Synthesis of the inputs and outputs in the sensitivity analysis.

Fig. 3b presents the case when cement hydraulic integrity is checked, meaning that the risk of having a microannulus propagate
upward from the bottom of the plug or downward from the top of the plug is simulated. This figure shows that a microannulus is
induced at the top of the cement plug, but it does not propagate throughout the length of the plug. The aperture of the cement plug at its
top is 217 mm. Fig. 3c presents the case when cement hydraulic integrity is checked after the WHP is increased to 10 MPa (1,450 psi),
indicating that a microannulus has been created throughout the length of the plug.

4 2019 SPE Drilling & Completion

ID: jaganm Time: 15:42 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 5 Total Pages: 11

Microannulus Width (μm) Microannulus Width (μm) Microannulus Width (μm)

–800 –400 0 400 800 –800 –400 0 400 800 –800 –400 0 400 800
700 700 700

750 750 750

800 800 800

Depth (mMD/RT)

Depth (mMD/RT)

Depth (mMD/RT)
850 850 850

900 900 900

950 950 950

Min Min Min
Med Med Med
Max Max Max
1000 1000 1000
(a) CMI after cement hydration (b) CHI after cement hydration (c) CHI during pressure testing

Fig. 3—Microannulus during the CMI and CHI simulations—base case.

The three curves, presented on each graph of Fig. 3, show the impact of the formation Young’s modulus. The minimum (Min) case
is when the formation Young’s modulus value is 3.75 GPa (0.544 Mpsi), the medium (Med) case is when the formation Young’s
modulus value is 5 GPa (0.725 Mpsi), and the maximum (Max) case occurs when the formation Young’s modulus value is 6.25 GPa
(0.907 Mpsi). The CMI simulation (Fig. 3a) leads to the conclusion that, when all other parameters are the same, the more flexible the
formation, the lower is the risk of opening a microannulus. Such an observation is confirmed by Fig. 3b, at the depths where the micro-
annulus is still closed (in the green zone). Conversely, when the microannulus is open (red zone in Figs. 3b and 3c), the stiffer is the for-
mation, and the smaller is the aperture of the microannulus. For example, this aperture at the top and the bottom of the cement plug is
684/597/544 and 138/117/109 mm for the Min/Med/Max cases, respectively, during the pressure test (Fig. 3c). This phenomenon can be
explained by the fact that, if the formation is stiff, the decrease in radial stress in the cement during hydration is fast, and it is difficult
to open and propagate a microannulus because the system is less compliant compared with a flexible formation. The induced pressure
decay during the pressure test is 2.35/1.40/0.98 kPa/min (0.34/0.20/0.14 psi/min), which might indicate that the pressure test would
have been judged successful.

Sensitivity Analysis. Next, in Figs. 4 through 6, the following convention is adopted: (1) dashed curve ¼ CMI after cement hydration;
(2) plain and thin curve ¼ CHI after cement hydration; and (3) plain and thick curve ¼ CHI during pressure testing.

Microannulus Width (μm) Microannulus Width (μm)

–800 –400 0 400 800 –800 –400 0 400 800
700 700

750 750

800 800
Depth (mMD/RT)

Depth (mMD/RT)

850 850

900 900

950 950

1000 1000
Openhole CMI Cased hole CMI E = 10 GPa CMI E = 7 GPa CMI
Openhole CHI before test Cased hole CHI before test E = 10 GPa CHI before test E = 7 GPa CHI before test
Openhole CHI after test Cased hole CHI after test E = 10 GPa CHI after test E = 7 GPa CHI after test

(a) Sensitivity on the hole type (b) Sensitivity on the cement Young’s modulus

Fig. 4—Sensitivity of the microannulus on the type of hole and the cement Young’s modulus (E ). Dashed curves 5 CMI after
cement hydration; plain curves 5 CHI after cement hydration (thin) and during pressure testing (thick).

2019 SPE Drilling & Completion 5

ID: jaganm Time: 15:42 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 6 Total Pages: 11

Microannulus Width (μm) Microannulus Width (μm)

–800 –400 0 400 800 –800 –400 0 400 800
700 700

750 750

800 800
Depth (mMD/RT)

Depth (mMD/RT)
850 850

900 900

950 950

1000 1000
OH = 8½ in. CMI OH = 12¼ in. CMI Q = 604 MJ/m3 CMI Q = 300 MJ/m3 CMI
OH = 8½ in. CHI before test OH = 12¼ in. CHI before test Q = 604 MJ/m3 CHI before test Q = 300 MJ/m3 CHI before test
OH = 8½ in. CHI after test OH = 12¼ in. CHI after test Q = 604 MJ/m3 CHI after test Q = 300 MJ/m3 CHI after test

(a) Sensitivity on the hole diameter (HD) (b) Sensitivity on cement heat of hydration (Q)

Fig. 5—Sensitivity of the microannulus on the hole diameter (HD) and the cement heat of hydration (Q). Dashed curves 5 CMI after
the cement hydration; plain curves 5 CHI after the cement hydration (thin) and during pressure testing (thick).

Microannulus Width (μm) Microannulus Width (μm)

–800 –400 0 400 800 –800 –400 0 400 800
700 700

750 750

800 800
Depth (mMD/RT)

Depth (mMD/RT)

850 850

900 900

950 950

1000 1000
Ty = 4.8 Pa CMI Ty = 9.6 Pa CMI k = 0.045 Pa.s CMI k = 0.090 Pa.s CMI
Ty = 4.8 Pa CHI before test Ty = 9.6 Pa CHI before test k = 0.045 Pa.s CHI before test k = 0.090 Pa.s CHI before test
Ty = 4.8 Pa CHI after test Ty = 9.6 Pa CHI after test k = 0.045 Pa.s CHI after test k = 0.090 Pa.s CHI after test

(a) Sensitivity on the fluid yield stress (b) Sensitivity on the fluid consistency

Fig. 6—Sensitivity of the microannulus on the fluid yield stress (sy) and consistency index (k). Dashed curves 5 CMI after cement
hydration; plain curves 5 CHI after cement hydration (thin) and during pressure testing (thick).

Fig. 4a presents a sensitivity analysis on the type of wellbore in which the plug is placed. The black curves are for a 0.216-m
(81=2-in.) open hole, while the red curves are for a 0.245-m (95=8-in.) casing cemented in a 0.311-m (121=4-in.) hole with a 7-GPa (1,015 psi)
Young’s modulus and a 0.2 Poisson’s ratio cement system. The figure shows that the cased hole configuration behaves similar to a
“stiff formation.”
Fig. 4b presents a sensitivity analysis on the cement Young’s modulus (E). The black curves are for a value of 10 GPa (1.450 Mpsi),
while the red curves are for a value of 7 GPa (1,015 psi). This figure shows that the impact of the cement Young’s modulus is very small
when a microannulus is not created. It also shows that the more flexible the cement system, the wider the microannulus, when created.
This phenomenon can be explained in the same manner as the impact of the formation stiffness: If the cement system is stiff, it is diffi-
cult to open and propagate a microannulus because the system is less compliant compared with a resilient cement system.

6 2019 SPE Drilling & Completion

ID: jaganm Time: 15:42 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 7 Total Pages: 11

Fig. 5a presents a sensitivity analysis on the HD. The black curves are for an HD value of 0.216 m (81=2 in.), while the red curves are
for a value of 0.311 m (121=4 in.). This figure shows that the larger the hole, the lower the risk of opening a microannulus, but the larger
the microannulus aperture, if it opens.
Fig. 5b presents a sensitivity analysis on the cement Q. The black curves are for a Q of 604 MJ/m3 (16,211 Btu/ft3), while the red
curves are for a value of 300 MJ/m3 (8,052 Btu/ft3). This figure shows that the impact of the cement Q is very small when a microannu-
lus is not created. It also shows that the greater the Q, the lower the microannulus aperture, if a microannulus is created.
Fig. 6a presents a sensitivity analysis on the fluid sy. The black curves are for a value of 4.8 Pa (10 lbf/100 ft2), while the red curves
are for a value of 9.6 Pa (20 lbf/100 ft2). This figure shows that the impact of the fluid sy is null when no microannulus is opened. It also
shows that the greater the sy, the smaller the microannulus aperture, if it opens. This phenomenon can be explained by the fact that if
the sy is high, more energy is required for the fluid to flow in the microannulus. Therefore, the microannulus has not propagated
throughout the cement-plug length in the case of the 9.6-Pa (20-lbf/100 ft2)sy, meaning that a pressure decline should not occur.
Fig. 6b presents a sensitivity analysis on the fluid-consistency index (k). The black curves are for a value of 0.045 Pas
(0.093 lbf-sec/100 ft2), while the red curves are for a value of 0.090 Pas (0.187 lbf-sec/100 ft2). This figure shows that the impact of the
fluid consistency is null when a microannulus is not created. It also shows that the consistency has a very low impact on the microannulus
aperture, if it opens. This phenomenon can be explained by a very low flow rate in the fracture, leading to a very low shear-strain rate.

Case Histories
The CHI model was used on four wells to simulate the hydraulic integrity of cement plugs set in wellbores and to model the effect of
the plug-pressure testing. The summary of the simulations (Table 2) shows that the CHI predictions of the pressure declines are close
to the pressure decays measured during the pressure tests.

Plug Pressure Test Pressure Decay

Predicted by CHI Measured
Casing OD MD/RKB Length WHP
Well Date (in.) (m) (m) Fluid Type (MPa) (kPa/min) (psi/min) (kPa/min) (psi/min)
1 Nov. 17 13⅜ 1214 220 NAF (1.42 sg) 7.6 10.3 1.5 5.5 0.8
2 Dec. 17 20 660 150 NAF (1.42 sg) 6.9 6.9 1.0 6.9 1.0
3 Oct. 17 7⅝ 6383 308 NAF (1.38 sg) 16.8 11.0 1.6 4.8 0.7
4 June 17 9⅝ 6828 902 Water (1.02 sg) 26.2 82.8 12 27.6 4.0

Table 2—Summary of the wells for which the CMI/CHI approach was used to predict the behavior of a cement plug submitted to a pressure
test. OD 5 outside diameter. MD 5 measured depth, RKB 5 rotary kelly bushing.

As an example, with Well 3, the simulations predicted the creation of a microannulus with a size of 519 mm at the top of the plug
during the pressure test to 16.83 MPa (2,440 psi) (Fig. 7) and an average pressure decay of 11 kPa/min (1.6 psi/min) (Table 2). The
plug on Well 3 was pressure tested with 16.83 MPa (2,440 psi), and the pressure decay was measured to 4.8 kPa/min (0.7 psi/min)
during the last 15 minutes of the pressure test (Fig. 8).

Depth CHI Microannulus Apertures (μm)

Well Architecture
(m/RKB) After Cement Hydration During Pressure Testing
6075 0 341 0 519

6383

Fig. 7—Well 3, outputs of the CHI simulator: microannulus size vs. plug depth before and during the 16.83-MPa (2,440-psi)
pressure test.

The simulations seem to overestimate the pressure decay during the pressure tests, with the highest difference being observed for the
test performed with water. The analysis indicated that two mechanisms could explain the differences:
• The simulations might not be able to correctly estimate the microannulus tortuosity, such as observed by Thérond et al. (2017)
and Skorpa and Vrålstad (2018). This would induce a difference between computed and measured-pressure decay that does not
depend on the type of fluid. Therefore, it would also be at the origin of the uncertainty observed for the three nonaqueous fluids
used for the first three pressure tests.

2019 SPE Drilling & Completion 7

ID: jaganm Time: 15:42 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 8 Total Pages: 11

Well-Pressure Test Well-Pressure Test—Last 15 Minutes

18 16.4

Pressure (MPa)
Pressure (MPa)
16
14 16.3
12
10 16.2
8
6
4 16.1
2
0 16
12:18:00 12:36:00 12:54:00 13:12:00 13:30:00 13:48:00 13:14 13:18 13:22 13:26 13:30 13:34
Time Time
(a) Complete pressure test (b) Pressure test decline period
(lasts 15 minutes, once compressibility is negligible)
showing a decrease of 75.8 kPa (11 psi) in 16 minutes

Fig. 8—Well 3, pressure vs. time during the pressure test: (a) complete pressure test, (b) pressure-test decline period (last
15 minutes, after compressibility is negligible), showing a decrease of 75.8 kPa (11 psi) in 16 minutes.

• If the wellbore fluid is a nonaqueous fluid, it might be difficult for it to penetrate the cement because of the existence of induced
capillary pressures and high fluid-loss control. On the contrary, if it is a water-based mud (WBM) or a brine, it can penetrate the
cement pores, especially if the fluid-loss control is not as good as for the nonaqueous fluid, which tends to increase cement pore
pressure, resulting in the swelling of the plug. Such a mechanism, observed in the laboratory by Opedal et al. (2018), could take
place at the top and the bottom of the cement plug, therefore locally improving the efficiency of the cement barrier, resembling a
champagne cork.
For a cement plug with 10-GPa Young’s modulus, 0.2 Poisson’s ratio, and 0.6 Biot’s coefficient, at a 6828-m total vertical depth
(TVD), with 1.02-g/cm3 SG water above it (Case 4 in Table 2), the maximum increase in pore pressure would lead to an expansion of the
plug by
ð1  2Þ  ð1 þ Þ ð1  2Þ  ð1 þ Þ ð1  2  0:2Þð1 þ 0:2Þ
Der ¼  b  DP ¼  b  ðq  g  TVDÞ ¼
E E 10; 000
1:02  9:81  6829 3
 0:6  ¼ 2:95  10 :                                              ð4Þ
1; 000
Such an expansion is able to close a microannulus at the boundary of a 95=8-in.-diameter cement plug equal to
OD 95=8  0:0254
 Der ¼  2:95  103 ¼ 360 lm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ð5Þ
2 2
The model used to simulate Cases 1 through 4 in Table 2 did not use this functionality. A modification of the CMI/CHI model has
been performed to include this functionality, and it is currently being validated.

Conclusions
The model for cement-plug integrity that has been developed includes two modules—a CMI module that simulates cement hydration to
obtain the state of stress and pore pressure throughout the cement plug, and a CHI module that checks if a microannulus can be propa-
gated in the same manner as a hydraulic fracture.
The examples presented in this document indicate that
• The pressure decay measured during a positive pressure test does not depend only on the sealing efficiency of the cement plug, but also
on the HD, the depth at which the plug has been placed, and the compressibility and viscosity model of the fluid that is above the plug.
• The most important risk for losing cement-plug integrity is through the creation of a microannulus at the cement-plug/hole interface.
The model also provides support that other risks, such as damaging the cement in compression or tension, are much less significant.
• Simply increasing the cement compressive and tensile strength does not always substantially improve the cement hydraulic-integrity
capability of the design.
• Cement tends to desaturate during hydration, except with the top and the bottom of the plug when the wellbore fluid can penetrate the
cement during hydration. This desaturation leads to a decrease of pore pressure in the cement plug and, according to Eq. 3, to a
decrease in stresses in the cement plug, which means that, at the end of hydration, a microannulus is close to being created.
• This decrease of stress is larger when the hole appears stiff, compared with when the plug is in front of an unconsolidated formation.
• At the end of hydration, the fluid pressure at the top of the plug is larger than the radial stress at the interface between the plug and
the hole, meaning that a microannulus could be propagated similar to a hydraulic fracture.
• If a microannulus is created, its aperture is larger when the cement is resilient, compared with when it is stiff, therefore leading to
larger pressure decays during a positive pressure test.
• If the wellbore fluid is a nonaqueous fluid, it might be difficult for it to penetrate the cement because of the existence of induced capil-
lary pressures and good fluid-loss control. On the contrary, if it is a WBM or a brine, it can easily penetrate the cement pores, which
tends to increase cement pore pressure, resulting in the swelling of the plug. Such a mechanism could take place at the top and the
bottom of the cement plug, therefore locally improving the efficiency of the cement barrier, similar to a champagne cork.
• Including a post-expansion additive in the cement system, such as MgO or salt (sodium chloride) types of additives, might help to
prevent the creation of a microannulus at the well/plug interface, or to close a microannulus that is already open. More research and
development should be dedicated to this technology to ensure its efficiency, especially when considering that its effects might occur
quite late in the life of the cement plug (i.e., after any pressure test has already been performed).

Nomenclature
A ¼ Al2O3, aluminum oxide
b ¼ Biot’s coefficient

8 2019 SPE Drilling & Completion

ID: jaganm Time: 15:43 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 9 Total Pages: 11

C ¼ CaO, calcium oxide

C ¼ thermal capacity
Ca(OH)2 ¼ Portlandite
C-S-H ¼ calcium silicates hydrates
C2S ¼ dicalcium silicate
C3A ¼ tricalcium Aluminate
C3S ¼ tricalcium silicate
C4AF ¼ calcium aluminoferrite
d ¼ differentiation operator
div ¼ divergence operator
e ¼ exponent function
e1 , e2 ¼ microannulus apertures
E ¼ Young’s modulus
Ea/R ¼ Arrhenius constant
F ¼ Fe2O3, Iron (III) oxide
G ¼ shear modulus
g ¼ standard acceleration caused by gravity
I ¼ unit tensor
Kd ¼ bulk modulus under drained condition
Ku ¼ bulk modulus under undrained condition
l ¼ liter
m ¼ fluid mass per unit of total initial cement pore volume (PV)
M ¼ Biot’s modulus
MgO ¼ magnesium oxide
pp ¼ pore pressure
PLR0 ¼ maximum-allowable pressure-loss rate
Q ¼ heat of hydration
q1 ¼ total heat produced when cement is completely hydrated
R ¼ gas constant in the ideal gas
s ¼ seconds
sd ¼ shrinkage under drained condition
su ¼ shrinkage under undrained condition
SG ¼ specific gravity
S ¼ SiO2, silicon dioxide
t ¼ time
tr ¼ trace operator
T ¼ temperature
 ¼ Poisson’s ratio
VLR0 ¼ maximum-allowable volume-loss rate
w/c ¼ water/cement ratio
DP ¼ increase in pressure in the microannulus
DPp ¼ increase in pore pressure in the cement plug
a-C2SH ¼ alpha dicalcium silicate hydrate
ad ¼ linear thermal-expansion coefficient under drained condition
au ¼ linear thermal-expansion coefficient under undrained condition
d ¼ increment operator
dpf ¼ increment of shear stress along the microannulus faces
e ¼ strain
k ¼ thermal conductivity
lA ¼ microannulus size
n ¼ hydration degree
r ¼ stress
q ¼ specific gravity
qf ¼ cement pore-fluid density
sy ¼ fluid-yield stress
~
AðnÞ ¼ normalized affinity

Acknowledgments
The authors thank BP, AkerBP, and CURISTEC for the permission to publish this paper. They also thank Caifeng Jin from CURISTEC
and Joaquı́n Reñe, Cesar Calvo, and Roberto Sánchez from CurisIT for their contribution in the development of the CMI/CHI applica-
tion. CURISTEC also acknowledges the contribution of SINTEF and the Research Council of Norway through the project titled Closing
the Gaps in CO2 Well Plugging (CO2 plug). This final version of the paper has benefited from valuable comments from the SPE Drilling
& Completion technical editors.

References
Agkun, H. and Daemen, J. J. K. 1994. Performance Assessment of Cement Grout Borehole Plugs in Basalt. Eng. Geol. 37 (2): 137–148. https://doi.org/
10.1016/0013-7952(94)90048-5.
Agkun, H. and Daemen, J. J. K. 1997. Analytical and Experimental Assessment of Mechanical Borehole Sealing Performance in Rock. Eng. Geol.
47 (3): 233–241. https://doi.org/10.1016/S00013-7952(97)00021-5.

2019 SPE Drilling & Completion 9

ID: jaganm Time: 15:43 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 10 Total Pages: 11

ASTM C204-18, Standard Test Methods for Fineness of Hydraulic Cement by Air-Permeability Apparatus. 2018. West Conshohocken, Pennsylvania:
ASTM International. https://doi.org/10.1520/C0204-18.
Bois, A.-P., Garnier, A., Rodot, F. et al. 2011. How to Prevent Loss of Zonal Isolation Through a Comprehensive Analysis of Microannulus Formation.
SPE Drill & Compl 26 (1): 13–31. SPE-124719-PA. https://doi.org/10.2118/124719-PA.
Bois, A.-P., Galdiolo, G., and Badalamenti, A. 2015. Taking Holistic Approach to Designing Cement Sheaths Helps Maintain Zonal Isolation for Life of
Well, Increases ROI. Drill Contr (November/December): 82–87.
Daccord, G., Guillot, D., and James, S. 2006. Remedial Cementing. In Well Cementing, second edition, ed. E. B. Nelson and D. Guillot, Chap. 14,
503–547. Sugar Land: Schlumberger.
Delabroy, L., Rodrigues, D., Straume, M. et al. 2017, Perforate, Wash, and Cement PWC Verification Process and an Industry Standard for
Barrier Acceptance Criteria. Presented at the SPE Bergen One-Day Seminar, Bergen, Norway, 5 April. SPE-185938-MS. https://doi.org/10.2118/
185938-MS.
Dormieux, L., Kondo, D., and Ulm, F.-J. 2006. Microporomechanics. Chichester: John Wiley and Sons.
Ferg, T. E., Lund, H.-J., Mueller, D. T. et al. 2011. Novel Approach to More Effective Plug and Abandonment Cementing Techniques. Presented at the
SPE Arctic and Extreme Environments Conference and Exhibition, Moscow, 18–20 October. SPE-148640-MS. https://doi.org/10.2118/148640-MS.
Guillot, D. 2006. Rheology and Flow of Well Cement Slurries. In Well Cementing, second edition, ed. E. B. Nelson and D. Guillot, Chap. 4, 93–142.
Sugar Land: Schlumberger.
Mainguy, M., Longuemare, P., Audibert, A. et al. 2007. Analyzing the Risk of Well Plug Failure After Abandonment. Oil & Gas Sci. and Techn. 63 (3):
311–324. https://doi.org/10.2516/ogst.2007026.
Mura, T. 1987. Micromechanics of Defects in Solids. second edition. Dordrecht: Martinus Nijhoff Publishers.
Nagelhout, A. C. G., Bosma, M. G. R., Mul, P. J. et al. 2010. Laboratory and Field Validation of a Sealant System for Critical Plug and Abandon Situa-
tions. SPE Drill & Compl 22 (3): 314–321. SPE-97347-PA. https://doi.org/10.2118/97347-PA.
Nelson, E. B. and Barlet-Gouédard, V. 2006. Thermal Cements. In Well Cementing, second edition, ed. E. B. Nelson and D. Guillot, Chap. 10, 319–341.
Sugar Land: Schlumberger.
Opedal, N., Corina, A. N., and Vrålstad, T. 2018. Laboratory Test on Cement Plug Integrity. Presented at the 37th ASME International Conference on
Ocean, Offshore & Artic Engineering, Madrid, Spain, 17–22 June. OMAE-2018-78347.
Powers, T. C. and Brownyard, T. L. 1947. Studies of Physical Properties of Hardened Portland Cement Paste. J. Am. Concr. Inst. 43: 101–132, 249–336,
469–505, 549–602, 669–712, 845–880, 933–992.
Skorpa, R. and Vrålstad, T. 2018. Visualization of Fluid Flow Through Cracks and Microannuli in Cement Sheaths. SPE J. 23 (4): 1067–1074. SPE-
180019-PA. https://doi.org/10.2118/180019-PA.
Thérond, E., Bois, A.-P., Whaley, K. et al. 2017. Large-Scale Testing and Modeling for Cement Zonal Isolation in Water-Injection Wells. SPE Drill &
Compl 32 (4): 290–300. SPE-181428-PA. https://doi.org/10.2118/181428-PA.

SI Metric Conversion Factors


F ( F–32)/1.8 ¼ C
Btu/ft337.25895 E03 ¼ MJ/m3
Mpsi**6.894757 Eþ00 ¼ GPa
US gal3.78541 E03 ¼ m3
ft3.048* E01 ¼ m
in.2.54* E02 ¼ m
lbf/100 ft247.88026 Eþ00 ¼ Pa
lbm4.535924 E01 ¼ kg
lbm/gal0.119826 Eþ00 ¼ g/cm3
psi6.894757 Eþ00 ¼ kPa
psi6.894757 E03 ¼ MPa
*
Conversion factor is exact.
**
See text for definition.

Axel-Pierre Bois is the founder and chief executive officer of CURISTEC, a consulting company that provides research and tech-
nical service in the areas of petroleum geomechanics, wellbore stability, wellbore integrity, well plug and abandonment, oilwell
cement design, and cement/rock testing. His current research interests include cement-sheath and cement-plug integrity, well-
bore strengthening, and thermo/chemo/mechanical modeling of setting cement. Bois has authored or coauthored more than
35 technical papers. He earned an engineering degree from École des Mines, Nancy, France, and a PhD degree in geome-
chanics from the University of Sherbrooke, Canada. Bois is an SPE member, and currently serves as a technical editor for SPE
Drilling & Completion. He has chaired two SPE international cementing technical workshops, the first one in Lyon, France (2014)
and the second one in Abu Dhabi (2018).
Manh-Huyen Vu is the director of the testing laboratory of CURISTEC in Pau, France. His research interests include cement integ-
rity, thermo/chemo/mechanical modeling of setting cement, gas migration modeling, and geomechanical simulations. Vu has
authored or coauthored more than ten technical papers and holds two patents. He earned a PhD degree in petroleum
cementing and geomechanics from École des Ponts ParisTech (France). Vu is a member of SPE.
Kim Noël is a simulation engineer within CURISTEC. His research interests include cement integrity and gas-migration modeling.
Noël earned an engineering degree from Polytech Paris-UMPC (France). He is a member of SPE.
Anthony Badalamenti is the business development manager for CURISTEC. He joined CURISTEC in 2015; before joining CURISTEC,
he worked for Halliburton from 1979 to 2013. Badalamenti’s current research interests include cement integrity, cement model-
ing, and well construction. He has authored or coauthored more than 15 technical papers and holds more than 30 US patents.
Badalamenti earned a BS degree in civil engineering from Tulane University and a BA degree in history/government from Nicholls
State University. He is a member of SPE.

10 2019 SPE Drilling & Completion

ID: jaganm Time: 15:43 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030

 DC191335 DOI: 10.2118/191335-PA Date: 31-January-19 Stage: Page: 11 Total Pages: 11

Laurent Delabroy is lead engineer of cementing and zonal isolation at AkerBP. He was a senior technical engineer of cementing
at BP for 5 years and previously he held different positions within Schlumberger. Delabroy’s current interests include zonal isola-
tion, drilling fluids, and cementing. He has authored one paper. Delabroy earned an engineering degree (Diplôme Grande
École) from the École supérieure de Chimie Industrielle de Lyon (France) and an MS degree from Rice University. He is a member
of SPE.
Emmanuel Thérond is a cementing technical specialist at BP. He has worked for 27 years in wells technical and operational posi-
tions in Europe, Africa, the Middle East, and the United States. Thérond’s current interests include zonal isolation and well integ-
rity, lost circulation, cementing procurement, and technology deployment. He has authored or coauthored eight technical
papers, and he holds nine patents to date. Thérond earned an engineering degree (Diplôme Grande École) from the École
Nationale Supérieure des Art et Industries de Strasbourg, France. He is a member of SPE.
Knut Hansen is senior technical cementing specialist at BP. Previously, he worked for 20 years in various technical and manage-
rial capacities for Schlumberger. Hansen earned an MS degree in mechanical engineering from the University of Hannover.

2019 SPE Drilling & Completion 11

ID: jaganm Time: 15:43 I Path: //chenas03/cenpro/ApplicationFiles/Journals/SA/SPE/DC##/Vol00000/180030/Comp/APPFile/SA-DC##180030