Well Cementing Book (285-395) PDF
Well Cementing Book (285-395) PDF
Well Cementing Book (285-395) PDF
Table 7-19. Compressive Strength and Shear-Bond Strength of Conventional and Foamed Cements†, ‡
Composition Density Compressive Strength Shear Bond Strength Ratio of Shear to
(lbm/gal [g/cm3]) (psi [MPa]) (psi [MPa]) Compressive Strength (%)
Class G 15.8 [1.90] 4,200 [29.0] 403 [2.8] 9.6
Thermal and electrical conductivity Table 7-20. Resistivity of Foamed and Conventional
Short et al. (1961) reported that foams have lower Cements†
thermal conductivity, because of the presence of gas Cement type Density Specific Resistivity‡
voids and the lower amount of solids. Nelson (1986) (lbm/gal [g/cm3]) (ohm-cm)
reported that the thermal conductivity of cement Conventional 15.7 [1.88] 1.25 × 104
systems is roughly proportional to slurry density, regard-
less of whether the cement was foamed. These data are Foamed 9.3 [1.11] 1.4 × 10 4
presented in Fig. 7-26. † From Smith et al. (1984). Reprinted with permission from World Oil.
Studies of the resistivity of foamed cement indicate ‡ ASTM D-257
0.9 5.8 6.7 7.5 8.3 9.1 10.0 10.8 11.6 12.5 13.3 14.1 15.0 15.8
0.8
0.7
Thermal 0.6
conductivity Conventional
0.5 low-density
⎛ BTU ⎞
⎜⎝ hr × ft × °F ⎟⎠ systems
0.4
Foamed cements and
0.3 microsphere systems
0.2
0.1
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
Cement density (g/cm3)
Fig. 7-26. Cement density/thermal conductivity relationship (from de Rozières and Ferrière, 1991). Reprinted with permission of SPE.
14.0 [1,680] 479 [3.35] 538,000 [3,759] 0.90 3,270 [22.9] 0.008
exhibit improved elasticity, ductility, and expansion Table 7-22. Mechanical Properties of Set Cements
properties over a wide density range—14.3 to 19.0 lbm/ Containing Rubber Particles†
gal [1,720 to 2,280 kg/m3]. Le Roy-Delage et al. (2000) Slurry Tensile Young’s TS/E Compressive
described cement systems containing from 30% to 100% Density Strength, Modulus, E (× 1,000) Strength
BWOC rubber particles in the 40/60 mesh range. As (lbm/gal TS (psi [MPa]) (psi [MPa])
shown in Table 7-22, the resulting set cements are [kg/m3]) (psi [MPa])
more flexible. 12.0 [1,440] 93 [0.65] 73,900 [516.7] 1.26 412 [2.88]
EPS cement systems (Section 7-9) have been formu-
13.1 [1,570] 140 [0.99] 137,800 [963.6] 1.04 819 [5.73]
lated using flexible particles as one of the components
(Le Roy-Delage et al., 2000). The particles have the 13.6 [1,630] 207 [1.45] 189,000 [1,320] 1.11 1,130 [7.93]
following characteristics:
14.1 [1,690] 237 [1.66] 240,000 [1,678] 1.01 1,782 [12.46]
■ Particle size: ≤500 μm
■ Young’s modulus: <5,000 MPa, preferably <2,000 MPa 15.2 [1,820] 390 [2.73] 460,900 [3,223] 0.86 2,890 [20.21]
■ Poisson’s ratio: >0.3. 16.4 [1,970] 323 [2.26] 864,000 [6,042] 0.76 3,934 [27.51]
† After Le Roy-Delage et al. (2000). Reprinted with permission of SPE.
Thermoplastics, like polyamide, polypropylene, and
polyethylene, or polymers, like styrene divinylbenzene
or styrene butadiene, are compatible with these perfor-
mance criteria. Because the specific gravities of these
materials fall between 0.9 and 1.2, they can also reduce the
cement system density. Table 7-23 presents mechanical-
properties data from set cements containing various
amounts of flexible particles.
Table 7-23. Mechanical Properties of EPS Set Cements Containing Flexible Particles†
Flexible Particle Slurry Density Tensile Strength, Young’s Modulus, E TS/E Compressive Strength,
(volume %) (lbm/gal [kg/m3]) TS (psi [MPa]) (psi [MPa]) (× 1,000) CS (psi [MPa])
Styrene divinylbenzene (25) 14.0 [1,680] 365 [2.52] 521,400 [3,595] 0.72 4,860 [33.5]
Styrene divinylbenzene (30) 12.1 [1,450] 160 [1.10] 194,200 [1,339] 0.84 1,930 [13.3]
Polyamide (25) 14.0 [1,680] 406 [2.80] 374,200 [2,580] 1.09 4,050 [27.9]
Polypropylene (19) 14.0 [1,680] 329 [2.27] 426,600 [2,941] 0.78 3,130 [21.6]
Polypropylene (24) 13.7 [1,640] 381 [2.63] 438,000 [3,020] 0.88 3,810 [26.3]
Styrene butadiene (25) 14.2 [1,700] 299 [2.06] 302,400 [2,085] 1.00 2,100 [14.5]
Polyethylene (25) 13.6 [1,630] 306 [2.11] 299,800 [2,067] 1.04 3,320 [22.9]
† After Le Roy-Delage et al. (2000). Reprinted with permission of SPE.
400
7-11.4 Fibers
Adding fibers or ribbons to a cement matrix also 200
improves flexural strength. Nylon fibers have been used 0
for many years for this purpose (Chapter 3). More 0 0.2 0.4 0.6 0.8 1
recently, Le Roy-Delage et al. (2000) and Baret et al. Deflection (mm)
(2002) described the addition of metallic microribbons
to improve impact resistance, toughness, and tensile Fig. 7-27. Typical load-deflection curve comparing neat and micro-
strength. The principal applications of this system are ribbon cement systems.
kickoff plugs and multilateral junctions.
The performance of the metallic ribbon system is
illustrated in Fig. 7-27, which shows a load-deflection
curve recorded during a flexural test. The plot shows the
amount of force required to bend or deflect the set-
cement sample a given distance. The results show that
a neat slurry fails completely after being deflected
less than 0.1 mm. The microribbon slurry was able to
bear a load after a deflection of nearly 1 mm.
160 13,000
140 12,000
120
11,000
100
Setting Compressive 10,000
time 80 strength
(min) (psi) 9,000
60
8,000
40
20 7,000
0 6,000
30 40 50 60 70 80 90 0 10 20 30 40 50
Temperature (°C) Time (hr)
Fig. 7-29. Setting characteristics of a magnesium potassium phosphate cement system (Wagh and Brown, 1999).
Expansive cements Improved cement/casing and Commercial expanding cements or Portland Formation of ettringite crystals
cement/formation bond cement containing calcium sulfate hemihydrate
Freeze-protected cements Cementing across permafrost zones Calcium aluminate cement Rapid strength development at
low temperatures
Salt cement systems Cementing across salt zones Portland cements containing sodium Systems do not disturb salt-bearing
or sensitive formations chloride or potassium chloride at formations
concentrations up to saturation Systems do not disturb sensitive clays
Cements for corrosive Cementing chemical waste Epoxy-based cement systems Chemically inert to strong acids and
environments disposal wells Elastomeric composites bases
Cementing CO2 injection wells Pozzolanic cement systems Reduced cement-matrix permeability,
BFS systems improved chemical resistance
BFS systems Alternative to or supplement for BFS + activator (e.g., sodium hydroxide Formation of C-S-H phase, with extensive
Portland cement or Portland cement) incorporation of Al, Mg, Fe, and sulfate
Conversion of drilling fluid to cement BFS + activator mixed with drilling fluid in structure
EPS cement systems Systems with improved Cement blends with multimodal Particle-size distribution minimizes mix-water
placement and set-cement particle-size distributions concentration required to prepare pumpable
properties over a wide slurry- slurry; set cement is less permeable than
density range conventional systems
Ultralow-density cement Cementing across formations Cements containing glass or ceramic Low slurry density reduces hydrostatic
systems with low fracture gradients or microspheres pressure in wellbore and prevents
that are vuggy or cavernous Foamed cements formation breakdown
Flexible cement systems Improved resistance to stresses Cements containing flexible particles Flexible particles decrease Young’s modulus
induced by perforating, hydraulic and increase Poisson’s ratio
fracturing, and tectonic movement
Microfine cement systems Squeeze cementing, sealing casing leaks Portland or BFS cements with small particle Smaller particles more readily enter cracks
size (4 to 15 μm) and high surface area and formation pores
(500 to 1,000 m2/kg)
Acid-soluble cement Temporary solution for lost circulation Magnesium oxychloride (Sorel) cements Principal binder phase is readily removed
systems by contact with a strong acid (e.g., HCl)
Chemically bonded Fast-setting cements that develop high Magnesium potassium phosphate Acid-base reaction between an acid
phosphate ceramics compressive strength phosphate and a metal oxide
Storable cement slurries Eliminates handling of dry powders Portland, BFS or blended cements During cement job, concentrated slurry
during cementing operations slurried in aqueous solution containing is diluted with mix water containing
a strong cement retarder activator (e.g., sodium silicate)
θ
σx
ΔA τxy x
σy
Fig. 8-1. Stress vector definition.
( ) ( ) ( )
which it acts. 1 ⎡ 2 1 2⎤
In a two-dimensional situation, if σx, σy , and τxy are σ 1 = σ x + σ y + ⎢ τ xy + σ x − σ y ⎥ (8-5a)
known (Fig. 8-2), the stress state on any plane oriented 2 ⎣ 4 ⎦
at an angle θ from σx can be expressed as follows:
and
( 2
)( )(
σ n = σ x cos θ − 2τ xy sin θ + σ y sin θ 2
) (8-2) 1/ 2
( ⎡
) ( ) ( )
1 2 1 2⎤
σ 2 = σ x + σ y − ⎢ τ xy + σ x − σ y ⎥ (8-5b)
and
2 ⎣ 4 ⎦
where θ is given by Eq. 8-4.
⎡1
( )⎤
(
τ = ⎢ σ y − σ x sin 2θ ⎥ + τ xy cos 2θ .
⎣2 ⎦
) (8-3) If one generalizes this concept to three dimensions,
it can be shown that six independent components of the
stress (three normal and three shear components) are
These expressions are obtained by writing equilib- needed to define the stress unambiguously. The stress
rium equations of forces along the σ n and τ directions, vector for any direction of Δ A can generally be found
respectively (see Fig. 8-2). Note also that the moment by writing equilibrium-of-force equations in various
equilibrium implies that τxy is equal to τyx. There are directions. There are three principal planes for which
always two perpendicular orientations of ΔA for which the shear-stress components vanish and three principal
8-2.2 Strain
8-3. Cement behavior
When a body is subjected to a stress field, the relative
positions of points within it are altered. The body A cement sample, like any material, deforms when sub-
deforms. If the new positions of the points do not corre- jected to stress. Determining a relationship between stress
spond to a translation and/or a rotation (i.e., by rigid-body and strain is an important aspect of solid mechanics. This
motion), the body is strained. This strain along an arbi- relationship is called the constitutive equation of the
trary direction can be decomposed into two components: material under consideration, and various theories have
been developed to describe it in a simplified way. The sim-
■ an elongation, defined as
plest one is the theory of elasticity, which assumes a
unique relationship between stress and strain (and that
L* − L the behavior is reversible). This theory is usually sufficient
ε = lim (8-6)
L→0 L to analyze cement failure in tension or in compression at
ambient conditions. Other theories, such as the theory of
■ and a shear strain, defined as elastoplasticity, have been developed to take into account
nonreversible behaviors that are observed in materials
γ = tan Ψ ,( ) (8-7) before failure. Significant nonreversible behavior is
observed in cements subjected to confining pressure.
where Ψ is the change of angle between the two direc-
tions that, before straining, were perpendicular (Fig. 8-3).
Consequently, strain (being either a ratio of lengths 8-3.1 Stress-strain curve
or a change of angle) is dimensionless. If one assumes Fig. 8-4 presents a typical stress-strain relationship for
that the stresses are positive in traction, a positive lon- cement. The test is carried out under constant confining
gitudinal strain, ε, corresponds to an increase in pressure and constant axial strain rate (Appendix B).
length. Just as in the case of stresses, principal strains The sample is protected from the confining fluid by an
can be defined as longitudinal strain components acting impermeable flexible jacket. Measurements include the
on planes in which the shear strains have vanished. axial stress, the axial strain, and the radial strain. When
The analogy between stress and strain analyses is a confining pressure is applied to the sample, the origin
not completely valid; equilibrium equations must be of the stress-strain plots is usually translated to remove
satisfied by the stresses and compatibility equations by the influence of the hydrostatic loading on the stress
the strains. These relationships place some restrictions and strain (i.e., the axial stress is actually the differ-
on the local variation of stress and strain in the neigh- ential σa – pcon), where σa is axial stress and pcon is the
confining pressure.
O x
F
P A Strength
C
B
P′ Axial stress
pcon (compression)
A′ E
y A
150 ∂2 T ∂2 T ∂2 T ρC ∂T
Temperature + + = × , (8-18)
(°C) ∂x 2
∂y2
∂z 2 K ∂t
100
50
where K is the thermal conductivity, ρ the density, and
C the specific heat. A temperature gradient will develop
0 and, as shown in Fig. 8-7, a heat front will move with
60 65 70 75 80 85 90 time.
Distance from well axis (mm)
Fig. 8-7. Temperature profile in the cement sheath at different 8-3.4 Poroelasticity
times, following a 200°C temperature increase in the wellbore: after The pore pressure can be expected to change in the
1 min. (blue), 10 min. (red), and 100 min (green).
cement sheath during the life of the well because, just
after the hydration and the resulting water consump-
tion, the cement pore pressure drops to a very low value
Large temperature changes can affect the mechanical (Chapter 9). Once the cement has set, fluid from the
properties of set cement through chemical transforma- formation will flow into the set cement to equilibrate
tions (see Chapter 10); however, the thermal dilation of the pressure, and the set-cement pore pressure will
the steel, cement, and rock can have a greater influence. increase. A rapid increase of wellbore pressure and
All of these materials expand as the temperature temperature will also lead to a pore-pressure increase.
increases and contract as the temperature decreases. Pore fluids in cement play an important role because
Damage may occur owing to nonuniform heating. A they support a portion of the total applied stress. Thus,
portion of the material that is being heated might be only the remainder of the total stress, the effective stress
prevented from expanding, while an unheated portion component, is carried by the cement matrix (Fig. 8-8).
might be subjected to forced expansion. This generates In 1923, Van Terzaghi first introduced the effective-
stresses in the set cement that may lead to failure or stress concept for one-dimensional consolidation of
debonding. To determine the thermoelastic behavior, soils, and proposed the following relationship:
such stresses must be added to those generated by
elasticity. Set cement, rocks, and tubulars will follow the σ ′ = σ + p pore , (8-19)
same behavior, although they will differ quantitatively.
The deformation caused by thermal effects in where σ is the total applied stress, σ′ is the effective
absence of stress is given by stress governing the consolidation of the material,
and ppore is the pore pressure (note that if compressive
ε = β T, (8-16) stresses are taken to be positive, the equation is
σ′ = σ – ppore). Biot (1941, 1956) later proposed a con-
where β is the coefficient of linear thermal expansion and sistent theory to account for the coupled diffusion and
Τ is the temperature. Therefore, for the stress-strain rela- deformation processes that are observed in elastic mate-
tionship: rials. For time-independent processes, this poroelastic
material behavior is similar to that of an elastic solid
σx ν
εx −βT = − σ +σz ,
E E y
( ) when the stresses in Eq. 8-14 are replaced by effective
stresses such that
σy ν
εy −βT = − σ +σz ,
E E x
( ) σ ′ = σ + αppore . (8-20)
)
τ = Yco − tan ( Φ σ n ′ , (8-29)
⎛ π Φ⎞
( )
σ 3 ′ = − σ c + tan 2 ⎜ + ⎟ σ 1′
⎝4 2⎠
(8-30)
0
Uniaxial
–10 17.24-MPa confinement
–30
–60
–70
–50,000 –40,000 –30,000 –20,000 –10,000 0 10,000 20,000
Microstrain
Fig. 8-12. Radial and axial strain as a function of axial differential stress (axial stress – confining stress)
and confining pressure. Stress and strain values follow the convention that tensile stress is positive
(i.e., compression negative). Conventional neat system.
a variety of temperature and pressure conditions, such a 8-4 Mechanical behavior of a cement
wide range of results is probably caused by the formation
of free water during the experiments (Justnes et al., cased wellbore
1996) and the failure to control the boundary conditions. 8-4.1 State of stress in the cement sheath
If bulk-volume variations are measured when the To determine whether a cement sheath will fail or debond
cement has access to additional water during the test in the annulus, one must calculate the state of stress.
(e.g., by measuring the dimensional variation of an The calculated stress is then entered into an expression
annular ring mold or cylindrical sleeve filled with the to determine whether failure is attained. To calculate
cement paste and placed in water), a bulk expansion as the state of stress, one must assume a deformation
high as 0.3% by volume is observed after the cement sets behavior (e.g., elasticity) and consider the various
(de Rozières and Sabins, 1995). Uncontrolled bulk applied loads at specific boundaries such as the
expansion can be as harmful as bulk shrinkage, because casing/cement and cement/formation interfaces. In some
it can disrupt the casing/cement interface (Beirute et cases, the influence of temperature and pore pressure
al., 1988; Baumgarte et al., 1999). To avoid this prob- must also be considered.
lem, one must ensure that the cement has a lower In recent years, various models have been devel-
Young’s modulus value than the surrounding rock oped to analyze the state of stress in the cement
(Baumgarte et al., 1999, Le Roy-Delage et al., 2000). (Thiercelin et al., 1997; Bosma et al., 1999; Gino di
In the context of the mechanical properties of well Lullo and Rae, 2000; Fleckenstein et al., 2000;
cements, the initial phase of shrinkage is not relevant Philippacopoulos and Berndt, 2002; Pattillo and
(Setter and Roy, 1978). During this phase, the cement is Christansen, 2002). They are based on analytical solu-
still a liquid slurry. During primary cementing, before tions, numerical solutions, or a combination of both.
the cement slurry begins to set, the top of the cement
column moves downward to compensate for the volume
reduction (Chenevert and Shrestha, 1991). The bulk 8-4.2 Modeling the cement sheath using
shrinkage that occurs after a rigid network of hydration thermoelasticity
products has formed and compressive strength begins to In this section the modeling of stresses in a cased
develop is relevant to the mechanical properties. wellbore containing a finite number of concentric
Previous studies of cement shrinkage have shown two casings is briefly described. A cross section of the well-
key behaviors that can be easily linked to porous elasto- bore is shown in Fig. 8-15. The stresses in the cement
plastic solid behavior controlled by effective stress are calculated assuming that casing, cement, and rock
(Thiercelin et al., 1998). When cement has free access to are thermoelastic materials. The casing/cement and
additional water, the external water flows into the cement/rock interfaces are also assumed to be either
cement pore space to compensate for the total chemical fully bonded or unbonded. Finally, it is assumed that the
shrinkage, and almost no bulk shrinkage is observed. In cement is under no internal “effective” stress after
some formulations, an expansion might even be observed. setting. This final assumption is obviously a strong
For this to occur, the cement permeability must be high
Cement
σr r
Casing
σr ν
)
Fig. 8-15. A cross section of the well.
εr −βT = − σ +σz
E E θ
(
σ ν
simplification. It is known that, after placement, the
cement slurry unloads as its gel strength develops
εθ − β T = θ − σ r + σ z
E E
( )
σ ν
)
(Chapter 9). Field observations (Cooke et al., 1983;
Morgan, 1989) tend to confirm that the total stress in ε z − β T = z − σ r + σθ
E E
(
the cement drops to at least the hydrostatic pressure
1
given by the saturating formation fluid (mainly water), γ rz = τ rz , (8-31)
justifying the zero-effective stress assumption. However, G
in some cases the pore pressure can drop below the
hydrostatic pressure, especially in a casing-to-casing where E, ν, and G are respectively the Young’s modu-
configuration. One can imagine a variety of situations, lus, Poisson’s ratio, and shear modulus, and β is the
depending on the cement properties, cementing coefficient of linear thermal expansion.
procedure, formation permeability, and nature of The temperature distribution as a function of time is
the saturating fluid. Nevertheless, in the absence of obtained from the heat diffusion equation, which is
better information, this simplification is appropriate. expressed under the assumption that the initial and
Consequently, to study the cement behavior, only the boundary conditions do not depend on θ, as
variations of pressure, stress, or temperature that occur
∂2 T 1 ∂ T ∂2 T ρ C ∂ T
after the cement sets are considered. + + = , (8-32)
The geometry of the problem is axisymmetric, with ∂r 2 r ∂r ∂z 2 λ ∂t
the axis of symmetry being the wellbore axis, allowing
the use of cylindrical coordinates r, θ, and z. The sim- where λ is the thermal conductivity, ρ the density, and
plest situation is when the boundary and initial condi- C the specific heat.
tions (wellbore and far-field states of stress and temper- Plane strain is also assumed, meaning that there is no
ature) are independent of θ. The variables of interest are axial movement. This is usually a good assumption,
then the radial displacement; radial stress, σr; tangential although axial movement could develop when casing
stress, σθ; axial stress, σ z; the shear stress, τrz; and the sections are being heated and axial casing deformation
temperature, T (which in practice is the temperature dif- is not prevented at the surface.
ference from a reference state). The tangential stress is The stress model uses analytical solutions that have
a principal stress. The radial and tangential stresses are been presented in Thiercelin et al. (1997). The solution
shown in Fig. 8-16. The sign convention is that tensile is constructed with the conditions that the radial
stresses are positive. Thermoelasticity provides a linear displacements and radial stress are continuous across
. the interface between two materials and the radial
relationship between the strains ε r, ε θ, ε z, and γrz,
stresses, and temperature, T. stress is compressive. If the radial stress is tensile (or
8-4.3 Influence of wellbore pressure increase Fig. 8-18. Tangential stress in the cement sheath as a function of
distance from the wellbore.
The most damaging wellbore pressure increases often
occur during a pressure test of the casing. It is indeed
unfortunate that, by checking the casing integrity, one
can damage the cement sheath. An increase of mud respectively. The wellbore pressure increase is 2,900 psi
weight, a hydraulic fracturing treatment, or a perfora- [20 MPa]. The openhole diameter is 7 in. [178 mm], the
tion test can also generate large wellbore pressure casing outside diameter (OD) is 5 in. [127 mm], and the
increases. A wellbore pressure increase induces a casing weight is 23.20 lbm/ft [34.53 kg/m]. Because we
radial elastic expansion of the casing, which in turn are using linear elasticity with a single loading condi-
loads the cement. tion, the stress values are a linear function of wellbore
The effect of a pressure increase on the state of stress pressure. Thus, doubling the wellbore pressure results in
in the cement sheath is shown in Figs. 8-17 and 8-18, in doubling the value of the stresses in the cement.
which the radial and tangential stresses in the cement These figures show that, in this case, the radial stress
are shown as a function of the distance from the well- is compressive and the tangential stress is tensile. The
bore axis. In this case, the Young’s moduli of the steel, tangential stress is about half the absolute value of the
cement, and rock are 29.0 × 106 psi [200 GPa], 0.725 radial stress. Because cements are about 10 times
× 10 6 psi [5 GPa], and 1.45 × 10 6 psi [10 GPa], weaker in tension than in compression, the cement
respectively. The Poisson’s ratios are 0.27, 0.15, and 0.2, failure will occur in tension. Tensile failure appears
when the tensile stress is greater than or equal to the
tensile strength. The highest value of the tangential
1.0 stress is at the steel/cement interface; therefore, this is
where failure should first occur. In this case, cement
0.5 failure will correspond to the initiation and propagation
of tensile radial cracks, because tensile cracks propagate
0 perpendicular to the direction of the maximum tensile
Radial stress. If the wellbore pressure increases by 2,900 psi
stress –0.5 [20 MPa], the value of the tangential stress at the
(MPa) steel/cement interface can be used to calculate the
–1.0 tensile strength the cement must have to avoid failure.
The cement Young’s modulus has a strong influence
–1.5
on the cement sheath response. This is demonstrated in
–2.0
Fig. 8-19, which shows the tangential stress as a function
60 65 70 75 80 85 90 of distance from the wellbore axis. In this case the
Young’s modulus of the cement is 0.725 × 10 5 psi
Distance from well axis (mm)
[500 MPa]. This time the tangential stress in the cement
Fig. 8-17. Radial stress in the cement sheath as a function of is less tensile, even compressive, near the cement/rock
distance from the wellbore. interface. This is because of the mechanical support pro-
Stress 2
(MPa) Table 8-4. Cement Failure as a Function of Cement
Properties During a Wellbore Temperature Increase
0
Formulation C D
–2 Tensile strength (psi [MPa]) 334 [2.3] 218 [1.5]
Well 1
Interzonal communication
9
Well 2
Gas to surface
ts
pl
en
p
m
m
Bu
Ce
Second, owing to a volume decrease, fluid loss may into the formation had been neglected. Baret (1988)
create space within the cement matrix that gas can confirmed the critical importance of fluid loss by more
occupy. Finally, fluid loss may be responsible for control- precise direct computations based on Darcy flow. He
ling the filtercake permeability, which ultimately influ- determined that, even in the presence of drilling mud
ences the migration path. Fluid-loss additives may also filtercake, API fluid-loss rates as low as 10 mL/30 min
act indirectly to reduce the permeability of the cement would sometimes be required to prevent annular bridging.
slurry (Chapter 6). It is important to mention that poor fluid-loss control
The importance of fluid loss as a contributing factor across permeable formations further up the hole can
to gas migration was first recognized by Carter and also impair full transmission of the hydrostatic pressure
Slagle (1970). At that time, the respective influences of to a gas zone. In 1976, Garcia and Clark observed gas
fluid-loss control and cement-slurry gelation were not migration when fluid loss occurred high in the hole,
fully understood. It was, however, pointed out that bridg- and hydrostatic pressure was no longer transmitted
ing or gelation owing to fluid loss could restrict the from the column above the bridging point to the bottom
transmission of hydrostatic pressure. In 1975, Christian of the hole.
et al. derived a method for calculating the fluid-loss con- Parcevaux (1987) discussed how fluid loss causes a
trol required to prevent bridging of the cement across pore-pressure decline and the formation of a void space
permeable formations during and after cement place- in the cement. The interstitial water in cement slurry is
ment. Christian et al. concluded that reducing the mobile; therefore, some degree of fluid loss always
American Petroleum Institute (API) fluid-loss rate to occurs when the annular hydrostatic pressure exceeds
less than 50 mL/30 min would reduce gas invasion and that of the formation. The process slows when a low-per-
lessen cement permeability. In 1977, Cook and meability filtercake forms against the formation wall
Cunningham described a procedure to analyze the gas and can stop altogether when the annular and formation
leakage potential based on a similar fluid-loss-rate com- pressures equilibrate. Once pressure equilibrium is
putation. However, Webster and Eikerts (1979) pointed attained, any volume change within the cement will pro-
out that, because earlier work was not based upon flow voke a sharp pore-pressure decline; consequently,
equations, the relative importance of fluid loss may because of the low cement compressibility, a void space
have been overemphasized. The positive influences of forms within the cement matrix, potentially inducing gas
the drilling mud filtercake and mud-particle invasion influx into the cement.
2,000 4,779 ft 6,659 ft interfaces (Section 9-3.3.7). Any or all of these processes
3,459 ft may contribute to the overall phenomenon of gas migra-
1,000 tion, and this limits the applicability of Eq. 9-2.
225 Geothermal Grachyov and Leonov (1969), Parcevaux (1987), and
Temperature 205 6,659 and 6,585 ft • 6,885 ft
Fig. 9-9. Annular gas flow test results (from Levine et al., 1979).
Reprinted with permission of SPE.
Fresh water
10
150°F bath
20
.. p1
. . Pressure
8.34-lbm/gal water
.. . .
. . . transducers and Depth
. . . .
thermocouples (ft)
Tubing
. .. . 30
. .. . .
. . .
. . ..
p2 16.4-lbm/gal cement
Porous plate 40
Meter
4.0 hr
3.0 hr
2.5 hr
2.0 hr
1.5 hr
1.0 hr
0.5 hr
0.2 hr
0 hr
50
Regulator
0 4 8 12 16 20 24 28
Nitrogen Pressure (psi)
Fig. 9-8. Schematic diagram of apparatus to measure hydrostatic pressure transmission of cement
slurries (from Levine et al., 1979). Reprinted with permission of SPE.
9-3.3.7 Microannulus
Tensile
Another path for gas to migrate is through a microan- ΔT failure
nulus—a gap that may form between the cement sheath
and the casing or the formation after the cement has set.
Compressional
A microannulus is a common occurrence that can result Radial stress failure
from any number of events during the life of a well. It has
long been known that a pressure decrease inside the
wellbore after the cement has set will result in a casing- Fig. 9-11. Tensile and compressional failure of the cement sheath
from wellbore stresses.
diameter reduction, leading to the formation of a
microannulus. This commonly occurs when the density
of fluid inside of the casing is reduced after the cement
job. A wellbore-temperature decrease will also reduce
the casing diameter. 9-4 Predicting short-term gas migration
An example of a situation in which both pressure and Gas migration is a complex physical phenomenon that
temperature reduction can occur is when the casing is comprises many facets; as a result, physical modeling of
closed in at the end of the cement job. The exotherm this phenomenon is a formidable problem. It is a non-
generated by cement hydration will cause thermal steady-state process involving changing pressure fields
expansion of the steel casing. In addition, fluids trapped and fluid saturations, and an evolving matrix structure.
inside the casing will heat up, causing further thermal
where
9-4.2.1 Formation factor
The first factor, the formation factor, is a dimensionless
dhole = diameter of the open hole (in.)
ratio of the reservoir’s productive capacity, kh, to the
dpipe = diameter of the pipe (in.) critical volume, Vcrit (Eq. 9-12). The critical volume is the
Kgfp = gas flow potential additional cement porosity that is created during setting
Kmpr = maximum pressure restriction by chemical shrinkage between the top of the gas zone
and the pressure balance point. A slurry porosity of 2% is
L = cement column length (ft)
assumed at this stage of hydration, and gas is assumed to
pob = overbalance pressure (psi). permeate the annulus in a uniform fashion over the
The GFP is a dimensionless number that can vary defined length. Assuming all other factors remain con-
between 0 and infinity, and the severity of the potential stant, as the value of the formation factor increases, the
gas migration problem is rated according to Table 9-2. risk of gas migration increases.
kh 467.7 khρslurry
fform = = , (9-12)
⎡
) − ( d pipe ) ⎥⎦
2⎤
Table 9-2. GFP Severity Ratings Vcrit
(
pob ⎢ dhole
⎣
2
⎡
)2 − ( d pipe ) ⎤⎥⎦ L
9-4.3.1 Formation parameter
(
2
0.05 π ⎢ dhole
The revised FP is based upon the reservoir’s ability to ⎣
deliver gas. To estimate the produced gas volume, Vgas, Vcrit = , (9-17)
4
across a unit section, the basic steady-state relation for
gases is used. where
dhole = hole diameter (m)
⎡
( ) − ( p ) ⎤⎥⎦
2 2
πkht ⎢ p pore ann
dpipe = pipe diameter (m)
Vgas = ⎣ , (9-15) L = length from top of gas zone to top of caprock (m)
⎛ r ⎞ Vcrit = critical annular volume
μ gas × pann ⎜ ln res + s⎟
⎝ rhole ⎠ 0.05 = the coefficient for space available gas in
the slurry.
where The FP can then be calculated from the ratio of the
h = gas zone length (m) produced gas volume to this critical annular volume by
k = formation permeability (m2) combining Eq. 9-15 and Eq. 9-17.
pann = annular hydrostatic pressure at top of gas Vgas
zone (Pa) K fp =
Vcrit
ppore = pore pressure at top of gas zone (Pa)
(
⎡
) ( ) 2⎤
2
rhole = hole radius (m) 80 kht ⎢ p pore − pann ⎥
rres = reservoir radius (m) = ⎣ ⎦
⎛ ⎞
s = dimensionless skin factor (s = 0 by default). ⎡
( ) ( ) ⎤ r
2 2
⎢ dhole − d pipe ⎥ Lμ gas × pann ⎜ ln res + s⎟
t = production time estimated from setting time (sec) ⎣ ⎦ ⎝ rhole ⎠
Vgas = gas volume, downhole conditions (m3)
μ gas = gas viscosity (Pa-s). (9-18)
The skin factor is highly dependent upon the base where
fluid of the drilling mud (oil or water). A typical skin Kfp = formation parameter.
value is equal to 20.
The gas viscosity, μgas, depends on pressure and tem- Various levels of severity may then be empirically
perature conditions. It is given by Eq. 9-16. This particu- assigned to the value of the FP (Table 9-4).
lar equation is for methane.
( )(
μ gas = 9.76 × 10 −6 + 0.0126 × 10 −6 T + )
{ ( )( )}
p ⎡⎢ 3.22 × 10 −9 − 4.84 × 10 −12 T ⎤⎥ ,
⎣ ⎦
(9-16)
>0.75 to 1 Critical
0 to 0.25 Low
i=0
j=0 j = nφ MD
2D Wellbore Concentration
Mesh Grid Versus Depth
9-4.3.2 Mud removal parameter
The second parameter is the MRP. With the advent of Fig. 9-12. Mathematical simulator for the MRP determination.
improved two-dimensional mathematical simulators for
fluid displacement (Chapter 5), the ability to quantify
the effectiveness of mud removal has advanced signifi- causing the annular pressure to drop below the hydro-
cantly in recent years. As shown in Fig. 9-12, such a static pressure. However, the fluid column has lower
mathematical simulator can be used to calculate cement pressure limits at all depths along the wellbore:
concentration on a mesh grid representing a wellbore. ■ pressure applied at the top of the annulus (normally
Perfect displacement would be predicted if a cement atmospheric pressure)
concentration value of 100% were indicated across the ■ pore pressure in front of permeable zones
entire length.
■ vapor pressure of water in front of impermeable zones
Using industry-accepted practices that have been
developed from field experience, positive zonal isolation (including casing-in-casing annuli).
will usually be achieved with 500 ft of cement coverage The pressure cannot drop below these limits at any
above the top of the gas-bearing zone. An MRP can thus depth. Gas migration can occur only when the annular
be calculated over this zone. pressure at a given depth drops to a value equal to or less
Dtog +500 ft
than the pore pressure of a gas-bearing zone at that
depth. The shear stress at the wellbore wall that causes
∫ ( Ycc ) dz ,
1
K mrp = (9-19) the pressure to reach this critical value for gas entry is
h Dbog the PDLP. The following equation uses oilfield units.
where
Dbog = Depth to bottom of gas zone K pdlp =
(
pob dhole − d pipe ), (9-20)
4L
Dtog = Depth to top of gas zone
h = length from bottom of gas zone to 500 ft above where
top of gas zone Kpdlp = pressure decay limit parameter
Kmrp = mud removal parameter L = the length of the cement column above the gas-bear-
Ycc = cement concentration value (calculated by ing formation
mathematical simulator). pob = overbalance pressure at the end of cement
placement, further defined as
Although the MRP is a useful design tool, one should
not forget the importance of the good cementing prac- pob = pmud + psp + pcem + pback − p pore , (9-21)
tices outlined in Chapter 5.
where
9-4.3.3 Pressure decay limit parameter pback = backpressure (i.e., atmospheric pressure + any
applied backpressure)
The third parameter is the PDLP, which is based upon
the concept of critical wall shear stress described by pcem = hydrostatic pressure from the cement column
Stiles (1997). When a fluid is placed inside a pipe or pmud = hydrostatic pressure from the mud column
annulus, a shear stress may be created along the wall,
10
CHP
9-5 Theoretical strategies for combating
short-term gas migration 1
Critical Upper time
Theoretically, the root causes for gas migration time boundary
(described in Section 9-3.1) can be combated by manag- Time
ing the annular pressure decline, reducing the space for
entry, and minimizing the path for migration. For short- Fig. 9-13. Plot of gel strength development versus time to define the
CHP.
term gas migration, these strategies must be addressed
during the postplacement period.
One concept for evaluating the postplacement period
is to plot the evolution of gel strength over time (Stiles, 10,000
1997). Because gel strength development tends to be
logarithmic, the log of gel strength can be plotted as a 1,000
straight line versus time (Fig. 9-13). The PDLP can be Impermeable matrix
calculated from Eq. 9-20 and plotted on the gel strength
graph so that the intersection of this value with the gel Gel strength 100
(lbf/100 ft2) PDLP
strength curve will define a critical time, tc, when gas
can first enter the annulus. An upper time boundary, tf,
10
represents the time beyond which the gel strength is too CHP
high to allow gas migration. This occurs after the cement
begins to set and become an impermeable matrix. tf is 1
plotted in the same manner as tc, by drawing a vertical Critical Upper time
line from the x-axis to the gel strength–development/ time boundary
impermeable-matrix gel strength intersection. The time Time
between tc and tf is the critical hydration period (CHP).
Fig. 9-14. Strategy for shortening the CHP by reducing the matrix
Stiles recognized three distinct strategies for permeability of the cement.
shortening the CHP and thus reducing the risk of gas
migration.
10 10
CHP CHP
1 1
Critical Upper time Critical Upper time
time boundary time boundary
Time Time
Fig. 9-15. Strategy for shortening the CHP by increasing the rate of Fig. 9-16. Strategy for shortening the CHP by increasing the PDLP.
gel strength development.
Finally, the CHP can be shortened significantly by 9-6 Practical solutions for combating
increasing the PDLP (Fig. 9-16). By reviewing Eq. 9-20, gas migration
which defines the PDLP, it can be seen that, unlike the
previous strategies, the PDLP cannot be affected by Practical strategies for combating gas migration can be
modifying any of the cement-slurry properties (with the classified according to the factors previously outlined in
exception of the density). Instead, the size of the annu- Table 9-1. Table 9-6 presents the strategies as a function
lar gap, the hydrostatic contributions of the fluids in the of the three critical root causes and the three time-
well, the length of the fluid column, and the relevant based categories for gas migration.
pressure boundaries control this parameter. Increasing
the size of the annular gap by decreasing the pipe size 9-6.1 Physical techniques
or increasing the openhole diameter is one way to
increase the PDLP. Incidentally, this would also result It has long been known that a number of physical tech-
in a reduction of the FP (Eq. 9-18), further reducing the niques can, under certain circumstances, help control
potential for gas migration. In most cases, however, nei- gas migration. These include applying annular backpres-
ther decreasing the casing diameter nor increasing the sure or small pressure pulses to the annulus, using exter-
openhole diameter would be considered economically nal casing packers (ECPs) and liner-top packers, and
viable options for managing gas migration. reducing the cement column height (including multi-
Increasing the overbalance pressure (Eq. 9-21) stage cementing). Such techniques are certainly valid
would also increase the PDLP. This could be achieved by under a variety of conditions, but well conditions often
increasing the density of any of the wellbore fluids or by limit their application.
increasing the length of a relatively higher-density fluid Application of annular backpressure after the cement
in the annulus. The overbalance pressure is also is in place increases the overbalance pressure exerted on
increased if backpressure is applied. It is important to gas zones, thus delaying the time when gas can enter the
remember that setting an annular packer will have the annulus. However, the presence of weak zones may
opposite effect. The packer will reduce the overbalance restrict this technique because of the risk of inducing lost
pressure and drastically reduce the PDLP. The final way circulation (Levine et al., 1979).
to improve the PDLP is to decrease the length of the Another technique for delaying gas entry, first
cement column; however, this would not be desirable described by Haberman and Wolhart in 1997, involves
because the effective length of the annular seal above applying pressure pulses to the annulus after the cement
the gas zone would be shorter. is in place. The pressure pulses are applied with com-
pressed air or water at approximately 100 psi at a fre-
quency of 30 to 60 sec/pulse. The concept behind this
technique is that the pressure pulses will disrupt gel
strength development in the cement and therefore main-
tain a hydrostatic overbalance for a longer period of time.
na na Compressible
cements Fig. 9-17. Use of ECPs (from Suman, 1984).
na na Expansive
cements
na na Flexible
annulus shortly after cement placement, thus reducing
cements the hydrostatic overbalance across gas zones below the
ECP. Slurry volume reduction below the packer, from
na na Mud removal fluid loss or chemical contraction, can then result in gas
† na = not applicable invasion of the cement in this interval at an even earlier
time. This could permit undesirable crossflow between
zones located below the packer.
ECPs (Fig. 9-17), inflated by mud or cement slurry, The technique of reducing the cement column height
control gas migration by forming a positive barrier in the stems originally from the work of Levine et al. (1979).
annulus (Suman, 1984). However, ECPs require a com- Viewing the mix-water hydrostatic-pressure gradient as
petent formation to seal against, and they complicate a natural step in the pressure reduction, they proposed
the job execution. Because of the small clearance minimizing the cement column height above the gas
between the uninflated ECP element and the borehole, zone using a very simple graphical method (Fig. 9-18).
such tools have been known to suffer mechanical The job would be designed such that the pressure sum of
damage while casing is being run or during circulation at an equivalent height of water plus the hydrostatic pres-
high rates. Also, it is not uncommon for the packers to sure above the cement would always exceed the forma-
set prematurely because of unexpected pressure fluctu- tion pressure. There is little doubt that this approach
ations during the course of the job. Parcevaux (1984a) can help the design process in a gross sense; e.g., severe
pointed out that ECPs can exacerbate some problems, risks of underbalance may be avoided. It has indeed
because they effectively isolate the lower portion of the been applied with success across some depleted sands,
7,000 used in the former Soviet Union (Kucyn et al., 1977) and
/ga
lw
Permeable or 45.7 cm
9-7.2 Small-scale gas migration testers
nonpermeable
section Various bench-scale or benchtop devices for characteriz-
ing gas migration have been described in the literature.
The first, described by Cheung and Beirute (1982), used
Check valve 1.0 m a modified API fluid-loss cell to investigate the hydro-
Hot-water jacket static pressure decrease and subsequent gas migration
Gas-entry line
from volume in a setting cement column (Fig. 9-24). Adaptations of
325-mesh screen measurement this model have recently been made commercially avail-
device
able and are known as gas migration analyzers or gas
Rubber diaphragm
Slurry 49.5 cm flow analyzers (Fig. 9-25). The test cell is a modified API
fill line HPHT fluid-loss cell with a hollow hydraulic piston at the
To pressure
recorder Thermocouple top of the cell. The piston is pressurized with mineral oil
To temperature recorder to simulate the hydrostatic overbalance. Fluid loss can
occur at both the top and the bottom of the cell, either
Fig. 9-22. Schematic diagram of test fixture used to study gas through standard 325-mesh fluid-loss screens or through
leakage (from Tinsley et al., 1979). Reprinted with permission of SPE.
actual formation cores. Once the cement slurry has
developed a predetermined gel strength, gas is injected
into the bottom of the cell at an appropriate differential
pressure. Gas flowmeters and pressure transducers mea-
sure any gas migration through the slurry. A separate gel
strength–development test (described in Section 9-7.3)
must be performed on the slurry before performing this
gas migration test.
Nitrogen gas
Top valve Backpressure receiver
Nitrogen gas
Pressure transducer
325-mesh screen
Slurry
To recorder
Nitrogen gas
Bottom valve Gas Pressure
flowmeter regulator
Figure 9-24. Gas flow simulator (from Cheung and Beirute, 1982). Reprinted with permission of SPE.
Cement
slurry
Water flow
Fig. 9-28. Cement hydration analyzer. Reprinted with permission from Elsevier.
100 100
80 95
Pressure
60 90
Pressure Temperature
(bar) Temperature (C°)
40 85
Gas valve
opened
20 80
Shrinkage
Gas flow rate
0 75
0 4 8 12 16 20 24
Time (hr)
100 100
80 95
60 90
Gas rate Temperature
Temperature
(scf/min) (C°)
40 85
Gas valve
20 80
opened
Shrinkage
0 75
0 4 8 12 16 20 24
Time (hr)
Fig. 9-29. Output from the CHA. Cement that is permeable to gas (top) will exhibit a stable pressure
in the cell equal to the gas-inlet pressure, and gas flow will be measured with the gas flowmeter,
while cement that is impermeable to gas (bottom) will exhibit a continued pressure decrease
caused by shrinkage and no gas flow will be measured.
10-1 Introduction
High-temperature wells present special cement system
design challenges. The physical and chemical behavior
Thermal Cements
10
40
4
1
30 3 1
Compressive 2 Water 2
strength 20 permeability 0.1
(MPa) 1 (mD) 3
10 0.01
4
0 0.001
0 1 0 1
Curing time (months) Curing time (months)
Fig. 10-1. Compressive strength and water permeability behavior of Portland cement at elevated
temperatures (from Nelson and Eilers, 1985).
300 Tricalcium
Truscottite Xonolite
Calcium hydroxide
Temperature silicate
(°C, log scale) γ-C2S hydrate hydrate
200
Gyrolite Hillebrandite
150
Z-phase 11 Å Tobermorite α-C2S hydrate
100 Afwillite
Hydrous
14 Å Tobermorite
silica
75
Ca(H3SiO4)2 C-S-H(I) C-S-H(II)
50
0 0.5 0.6 0.7 0.8 1.0 1.3 1.5 2.0 2.5 3.0
CaO/SiO2 mole ratio of starting material
Fig. 10-2. Formation conditions for various calcium silicates (from Taylor, 1964).
Reprinted with permission from Elsevier.
40 0.1
Compressive Water
strength permeability
(MPa) (mD)
20 00.01
0 0.001
0 1 3 6 12 24 0 1 3 6 12 24
Fig. 10-3. Compressive strength and permeability behavior of 16.0-lbm/gal Class G systems stabilized
with 35% silica (from Nelson and Eilers, 1985).
At 480°F [250°C] the phase truscottite (C7S12H3) The discussion so far has been limited to the behavior
begins to appear. As the curing temperature approaches of the silicate hydration products. To the authors’ knowl-
750°F [400°C], both xonotlite and truscottite are near edge, the hydrothermal behavior of the aluminates and
their maximum stable temperatures, and dehydration of aluminoferrites has not been specifically described in
the residual CH to C occurs. At higher temperatures, the the literature. The common hydrated aluminate and alu-
xonotlite and truscottite dehydrate, resulting in the dis- minoferrite hydrates (Chapter 2) are not typically
integration of the set cement. observed when Portland cements are cured hydrother-
In addition to the compounds cited above, other mally. Ettringite is not stable in hydrothermal condi-
–
phases such as pectolite (NC4S6H), scawtite (C7S6CH2), tions, and is not normally detected. Some of the Al3+,
reyerite (KC14S24H5), kilchoanite (C3S2H approxi- Fe3+, and SO42– ions from ettringite are incorporated
mately), and calciochondrodite (C5S2H approximately) into the silicate phases.
may appear in Portland cement systems cured at ele- The preceding discussion illustrates the complexity of
vated temperatures. These phases can affect the perfor- the hydrothermal behavior of calcium silicate hydrates.
mance of the set cement, even when present in small The performance of the set cement depends not only on
quantities. the downhole temperature, but also on the presence of
Cements containing significant amounts of truscot- subterranean brines and other minerals. As a result, the
tite are usually characterized by low permeability standard conditions for equilibrium transformations
(Gallus et al., 1978). The formation of pectolite, a that are reported in the literature are not always
sodium calcium silicate hydrate, is accompanied by observed downhole (Langton et al., 1980). Therefore,
cement expansion (Nelson and Eilers, 1982); in addition, the set cement must be considered to be metastable,
pectolite appears to render cements more resistant to because its composition can evolve as downhole condi-
corrosion by highly saline brines (Nelson and Kalousek, tions change.
1977; Nelson et al., 1981). Scawtite has been shown to
enhance cement compressive strength when present in
minor amounts (Eilers et al., 1983; Bell et al., 1989). In
general, set cements that consist predominantly of cal-
cium silicate hydrates with C/S ratios less than or equal
to 1.0 tend to have higher compressive strengths and
lower water permeabilities.
Si
100 Anorthite system (1.49 g/cm3)
10 Class G cement
90 Anorthite
20 Xonotlite
80 Garnet
30 Kilchoanite/afwillite
70 Pectolite
40 Prehnite
60 Wairakite
50 Epidote
50 Quartz
60 Diopside
40 Hauyne
70 Margarite
30
80
20
90
10
100
100 90 80 70 60 50 40 30 20 10
Al Ca
Fig. 10-6. Ternary diagram showing compositions of minerals in the CaO • Al2O3 • SiO2 system.
Table 10–2. Composition and Performance of Calcium Phosphate Cement Systems Cured at 600°F †
Water Calcium Sodium Fly Ash Foaming Foam Slurry Density 28-Day
(wt%) Aluminate Polyphosphate (wt%) Agent Stabilizer (lbm/gal Compressive
(wt%) (wt%) (wt%) (wt%) [kg/cm3]) Strength
(psi [MPa])
23.3 17.5 15.6 40.8 1.9 0.9 12.1 [1,450] 570 [4.0]
22.3 21.0 14.9 39.2 1.8 0.8 15.1 [1,810] 1,060 [7.4]
† from Brothers et al. (1999).
18 D 25 D 32 D
Calcium phosphate 16.0 500 0 0 0
10-7 Deep oil and gas wells 10-7.1 Thickening time and initial compressive
Wells with depths exceeding 15,000 ft [4,570 m], with strength development
bottomhole temperatures above 230°F [110°C], are Cement slurries for deep wells are usually designed to
common throughout the world. Since the 1970s, hun- have at least 3 to 4 hr of pumping time. However, there
dreds of wells with depths exceeding 25,000 ft [7,600 m] are several complicating factors that must be men-
have been completed (Arnold, 1980; Wooley et al., 1984). tioned.
Such wells represent a large investment of time and As the length of the casing string or liner increases,
money; therefore, obtaining a successful well completion the problem of achieving a cement seal becomes more
is of paramount importance. difficult (Suman and Ellis, 1977). In many cases, the
The procedures for cementing deep wells are basi- static-temperature differential between the top and
cally the same as those for shallower wells; however, bottom of the cement column can exceed 100°F [38°C].
because of the severe well conditions and more complex Sufficient retarder must be added to the cement slurry
well architecture, such wells are usually considered to to allow adequate placement time at the BHCT; conse-
be critical (Smith, 1987). Higher temperatures, nar- quently, such a slurry may be over-retarded at the top of
rower annuli, overpressured zones, and corrosive fluids the cement column, resulting in a very long waiting-on-
are commonly encountered. Consequently, the cement cement (WOC) time. If high-pressure gas exists behind
system design can also be complex, involving an elabo- the casing string or liner, the risk of gas invasion into the
rate array of retarders, fluid-loss additives, dispersants, cement is high (Chapter 9). In recent years, advances
silica, and weighting materials. One must be certain that have been made in retarder chemistry and cement-
the cement system can be properly placed and will main- system design that have helped to mitigate such prob-
tain zonal isolation throughout the life of the well. At lems (Chapter 3).
present, Portland cement is used in virtually all deep oil When designing cement slurries for deep, hot wells, it
and gas well completions. is very important to use accurate static and circulating
Typical casing programs and cementing procedures temperature information. Such data may be obtained
for deep wells are given in Chapter 13. Detailed informa- from drillstem tests, logs, special temperature recording
tion regarding cement additives is found in Chapter 3. In subs, or circulating temperature probes run during hole
this section, the design of appropriate cement systems conditioning (Jones, 1986). Computer simulators have
for deep high-temperature wells is presented. also been developed to better predict well temperatures
(Chapter 12). If fluids are circulated in the well for sev-
eral hours before cementing, the well temperature may
be lowered significantly. In such cases, one must be
careful not to overestimate circulating temperature and
over-retard the cement slurry.
50 50
40 40
1
1
30 Compressive 30
Compressive
4 strength
strength 2
(MPa) 20 (MPa)
20
2
3
10 10
3 4
0 0
0.03 1 3 6 12 24 0.03 1 3 6 12 24
Curing time (months) Curing time (months)
100 100
10 10
1
1 1
Water Water
permeability 3 permeability 3
(mD) 2 (mD) 0.1
0.1
1 4 2
0.01 0.01
4
0.001 0.001
0.03 1 3 6 12 24 0.03 1 3 6 12 24
Curing time (months) Curing time (months)
Fig. 10-8. Compressive strength and permeability performance of Fig. 10-9. Compressive strength and permeability performance of
conventionally extended Portland cement slurries—450°F [232°C] conventionally extended Portland cement slurries—600°F [315°C]
after Nelson and Eilers, 1985). Reprinted with permission from the (after Nelson and Eilers, 1985). Reprinted with permission from the
Petroleum Society of CIM. Petroleum Society of CIM.
System 1 contained Type F fly ash as an extender and Systems 2 and 3 were extended with perlite and ben-
was the heaviest of the four. Despite the density advan- tonite. System 2 performed well at both 450° and 600°F
tage and the highest initial compressive strength, the [232° and 315°C] with regard to compressive strength.
performance of System 1 over a 2-year period was The permeability of System 2 varied back and forth
no better than that of lower-density systems at 450°F across the 0.1-mD line. System 3 was the least dense of
[232°C], and was the poorest of the four at 600°F the four. The compressive strength performance was
[315°C]. This delayed degradation of fly-ash-containing adequate at both curing temperatures, but the perme-
systems was probably the result of alkali contaminants abilities were too high. It is important to point out that
in the fly ash. Such contaminants can slowly react and perlite is compressible, and its extending effect
form substituted calcium silicate hydrates, notably decreases as the hydrostatic pressure in the well
reyerite, with deleterious effects (Eilers and Root, 1974). increases (Chapter 3). For this reason, perlite is rarely
It is important to mention that cement degradation asso- used today. System 4, containing diatomaceous earth,
ciated with fly ash has not been observed at curing tem- was a rather poor performer in the strength category, yet
peratures below 450°F [232°C]. had low permeability.
Production well
Fig. 10-11. Geothermal power plant (from Geothermal Education Office, 2001). Drawing courtesy
of Geothermal Education Office, Tiburon, California, USA.
10-8.1 Well conditions associated with An economical geothermal reservoir requires that
geothermal wells large quantities of hot water or steam must be produced
With the exception of hot, dry rock completions with cir- from each well. Therefore, the reservoirs are usually nat-
culating temperatures as high as 500°F [260°C] (Carden urally fractured and have effective permeabilities that
et al., 1983, Duchane, 1994), most geothermal wells are are probably greater than 1 D. The integrity of the for-
not cemented under “geothermal” conditions, because mations ranges from poorly consolidated to highly frac-
the fluids circulated during drilling cool the formation. tured, and the fracture gradients tend to be low.
The maximum circulating temperatures during the Consequently, lost circulation is the most serious obsta-
cement job seldom exceed 240°F [116°C]; therefore, the cle to successfully cementing geothermal wells (per-
design of cement systems with adequate thickening sonal communication, Weber, 2003). It is not uncommon
times is usually not a problem. Most geothermal wells to have losses in the casing strings set above the target
are less than 10,000 ft [3,050 m] in depth. Downhole reservoir, and in many cases total losses occur before the
pressures are seldom above the water gradient. intended setting point for the intermediate string. For
The drilling programs for geothermal wells usually these reasons, low-density cement systems are required
call for setting surface and production casing above the by most geothermal operators (Nelson et al., 1981).
reservoir. In some cases, a slotted liner is hung through Lost circulation also hampers the determination of
the producing zone, but cementing the liner is not con- cement placement temperatures. Placement tempera-
sidered critical. It is very important to cement the cas- ture simulation and modeling is essential to formulate
ings to the surface; otherwise, creep or elongation will the appropriate cement system (Chapter 12).
occur because of thermal expansion when the well is
brought into production (Shryock, 1984).
1 4 10 40 100 400
10-8.2 Performance requirements and design
Average silica size (μm)
considerations
Geothermal wells arguably present the most severe con- 300°F [150°C]
ditions to which well cements are exposed. As a result, 450°F [232°C]
617°F [325°C]
the performance requirements are among the most
stringent. Geothermal well cements are usually designed
to provide at least 1,000 psi [7.0 MPa] compressive
strength, and no more than 0.1-mD water permeability Mesh 325 140 70
100
(API Task Group on Cements for Geothermal Wells, 80
1985). In addition, the set cement often must be resis-
Crystalline 60
tant to degradation by saline brines. composition
Silica-stabilized Portland cement compositions are 40
(%)
almost exclusively used to complete geothermal wells; 20
however, their dominance is being challenged by sys- 0
tems that offer better resistance to the severe chemical 20 40 60 80 100 120 140 160 175
environments. Each is described in this section. Average silica size (μm)
Scawtite Kilchoanite
10-8.2.1 Portland cement–based geothermal well Xonotlite Quartz
cement compositions
When Portland cement–based cement systems are
Mesh 325 140 50
expected to contact highly saline and corrosive geother- 10.0
mal brines, the particle size of the added silica is an 4.0
important consideration. As explained in Chapter 3, 1.0
there are two forms of silica commonly used in well Water 0.4
permeability 0.1
cementing: silica sand, with a particle size of approxi- 0.04
(mD)
mately 175–200 μm, and silica flour, with an average par- 0.01
ticle size of approximately 15 μm. Field personnel usu- 0.004
ally prefer silica sand, because its lower surface area 0.001
facilitates easier slurry mixing. However, in certain geot- 1 4 10 40 100 400
hermal environments, silica sand cannot be relied upon Average silica size (μm)
to provide adequate stabilization.
Eilers and Nelson (1979) investigated the effect of Fig. 10-12. Effect of silica particle size on the performance of Class
G cement cured in geothermal brine (from Eilers and Nelson, 1979).
silica particle size on the performance of Class G cement Reprinted with permission of SPE.
formulations cured at various temperatures in a geo-
thermal brine. The salinity of the brine was 25,000 mg/L
total dissolved solids. Figure 10-12 shows the relation-
ships between the silica particle size and several para- Figure 10-13 shows that the silica particle-size effect is
meters—compressive strength, water permeability, and significantly more pronounced with lower-density
cement phase composition. The slurry density was cement compositions.
15.8 lbm/gal [1.90 g/cm3]. A decrease in compressive High concentrations of sodium chloride depress the
strength and an increase in water permeability occurred rate at which silica enters solution (personal communi-
when the average particle size of the added silica cation, R. Fournier, 1979); as a result, when the silica
exceeded about 15 μm. Xonotlite was also replaced particle size is large, the rate of dissolution of silica is
by kilchoanite as the predominant cement phase. insufficient to allow the formation of the desired calcium
Fig. 10-13. Effect of silica particle size on the performance of a 13.5-lbm/gal Class G-perlite-bentonite
system cured in geothermal brine (from Eilers and Nelson, 1979). Reprinted with permission of SPE.
silicate hydrates (C/S ratio < 1). The kinetics of dissolu- 40% SiO2
tion can be affected by the particle size of the solute. 70 100% silica fume
Reducing the particle size of the silica increases its sur- 100% silica flour
60 50% silica fume, 50% silica flour
face area; consequently, a sufficient supply of silica is 33% silica flour, 67% silica fume
available. 50
25% silica flour, 75% silica fume
Grabowski and Gillott (1989) and Dillenbeck et al. Compressive
(1990) studied the effects of silica “fume,” with an aver- strength 40
age particle size of approximately 0.1 μm (Chapter 3), (MPa)
upon Portland cement systems at elevated temperatures 30
and pressures. With a constant SiO2 concentration (40%
20
BWOC) and water-to-solids ratio (0.5), samples were
prepared containing silica fume, combinations of silica 10
fume and silica flour, and silica flour. Curing was per- 7 28 56 90 210 270
formed at 450°F [230°C] and 400 psi [2.75 MPa] for
Total age (days)
7 days, using samples aged under ambient conditions for
periods up to 270 days. The systems containing silica 10–1
4,000
20
2,000
1,000 psi
0 0
Compressive strength of cement cube and sandstone cup samples after aging periods of 1 day and 3, 6,
and 12 months. Cup samples were cured and aged downhole. Cubes were laboratory cured under water
at 392°F [200°C] for 1 day, then exposed downhole for 3, 6, and 12 months in the Cerro Prieto geothermal
field, Mexico. The downhole temperature was 417°F [214°C].
Fig. 10-15. Compressive strength performance of typical geothermal well cements under
actual conditions (from API Task Force on Geothermal Well Cements, 1985). Reprinted
with permission from Oil & Gas Journal.
1 2 3 4 5 6 7 8 9 10
2 5
1 day
3 months
1 4
6 months
12 months
0 3
0.1 mD
Log10 –1 2 Log10
permeability permeability
(log mD) –2 1 (log nm2)
–3 0
–4 –1
–5 –2
–6 –3
Water permeabilities of cement samples taken from slurry-filled sandstone cup holders after
curing 1 day and 3, 6, and 12 months downhole in the Cerro Prieto geothermal field, Mexico.
The downhole temperature was 417°F [214°C].
Fig. 10-16. Permeability performance of typical geothermal well cements under actual
conditions (from API Task Force on Geothermal Well Cements, 1985). Reprinted with
permission from Oil & Gas Journal.
0.8
Steam
0.7
Oil
0.6
Thermal
0.5
conductivity
BTU/hr ft °F 0.4
0.3
0.2
0.1
5.8 7.5 9.1 10.8 12.5 14.1 15.8
Cement density (lbm/gal)
4
Compressive
strength 3
(thousand
psi)
2
100
10
Water
permeability 1
(mD)
0.1
0.01
1 3 6 12 24 1 3 6 12 24
Time (months) Time (months)
Fig. 10-19. Long-term performance of glass microsphere systems cured at elevated temperatures.
Typical slurries using glass or ceramic microspheres peratures (unpublished data, Nelson, 1987). X-ray dif-
are prepared with a silica-stabilized Portland cement– fraction analysis of the systems revealed the coincident
based slurry. The long-term performance of glass micro- appearance of reyerite and certain aluminosilicate
sphere systems cured at 450° and 600°F [232° and 315°C] hydrate phases. Ceramic microspheres are derived from
is shown in Fig. 10-19. The slurry densities vary from 10.0 fly ashes, and the delayed (reyerite-related) deteriora-
to 12.0 lbm/gal [1.20 to 1.45 g/cm3]. tion of normal-density fly ash cement systems has been
The performance of silica-stabilized ceramic micros- discussed earlier in this chapter.
phere systems at 450° and 600°F [232° and 315°C] is Typical foamed cement systems for thermal wells are
shown in Fig. 10-20. Initially, these systems were gener- prepared from a normal-density base slurry of Portland
ally stronger and less permeable than their glass micros- cement, at least 35% silica flour, a surfactant, and a
phere counterparts. However, between 1 and 2 years of foam stabilizer. The long-term performance at 450° and
curing, significant deterioration was noted at both tem- 600°F [232° and 315°C] of three foamed cement systems
4
Compressive
strength 3
(thousand
psi)
2
100
10
Water
permeability 1
(mD)
0.1
0.01
1 3 6 12 24 1 3 6 12 24
Time (months) Time (months)
Fig. 10-20. Long-term performance of ceramic microsphere systems cured at elevated temperatures.
with densities ranging from 9.0 to 12.0 lbm/gal [1.08 to Febus, 1984). Compressive strength and permeability
1.44 g/cm3] is shown in Fig. 10-21. Comparison of the data for systems cycled between 550° and 100°F [288°
foamed cement data with those of equal-density micros- and 94°C] are shown in Table 10-5. More recently, ther-
phere systems reveals the foams to have significantly mal cements containing additives that impart flexibility
higher compressive strength. The water permeabilities (Chapter 7) have been successfully introduced for
of the foamed cements are also higher (>0.1 mD), and steamflood applications (Stiles and Hollies, 2002; Stiles,
more variable with curing time. 2006).
Foamed cements have also been shown to resist
repetitive thermal cycling, which occurs when the cyclic
steam stimulation technique is applied (Harms and
4
Compressive
strength 3
(thousand
psi)
2
100
10
Water
permeability 1
(mD)
0.1
0.01
1 3 6 12 24 1 3 6 12 24
Time (months) Time (months)
Fig. 10-21. Long-term performance of foamed cement systems cured at elevated temperatures.
Table 10-5. Effect of Thermal Cycling on Performance of 10-9.2 In situ combustion wells
Foamed Cements for Steamflood Conditions† In situ combustion recovery, or fireflood, consists of ini-
Properties of Foamed Cement Density tiating combustion in an injection well and then propa-
Foamed Cement gating the combustion front by the injection of air or
10 lbm/gal 11.5 lbm/gal 13 lbm/gal
oxygen through the reservoir to the production wells
Compressive strength 1,210 psi 1,680 psi 2,260 psi (Chu, 1981; Petit et al., 1992). In such wells, the cement
after 20 days at 550°F
is exposed to maximum temperatures between 700° and
Compressive strength 1,630 psi 1,550 psi 2,440 psi 1,700°F [371°and 926°C] near the burning zone. Such
after 100 days at 550°F‡ temperatures exceed the stable range of Portland
Compressive strength 1,240 psi 2,020 psi 2,430 psi
cement; therefore, high-alumina cement is necessary.
after 160 days at 550°F§ Fireflood wells are physically similar to and are usu-
ally found in the same locations as steam injection wells.
Air permeability after 2.4 mD 1.0 mD 0.9 mD Thus, the formation conditions and cement performance
100 days requirements are basically the same. Usually, most of the
† Surface slurry: 15.4 lbm/gal Class G, 40% silica flour, 3% lime (from Harms and Febus, 1984).
Reprinted with permission of SPE. casing is cemented with Portland cement systems, with
‡ Cycled to 100°F twice.
calcium aluminate cement placed opposite and about
§ Cycled to 100°F three times.
538°C 538°C
38°C 93°C 315°C 815°C 38°C 93°C 315°C 815°C
15 100
10
10
1
Compressive Water
strength permeability
(MPa) (mD)
0.1
5
0.01
0 0.001
1 3 7 1 7 3 1 7
Curing time (days) Curing time (days)
Fig. 10-22. Compressive strength and permeability performance of calcium aluminate cement systems
at various temperatures (from Nelson and Eilers, 1985). Reprinted with permission from the Petroleum
Society of CIM.
10 1
Compressive Water
strength permeability
(MPa) (mD)
5 0.1
0 0.01
7 28 7 28
Curing time (days) Curing time (days)
Fig. 10-23. Performance of foamed calcium aluminate cement systems at 1,250°F [677°C] (from Nelson
and Eilers, 1985). Reprinted with permission from the Petroleum Society of CIM.
■
10-10 Conclusion Microsphere cement systems can be used in thermal
The preceding discussion has demonstrated that ther- wells, provided the base slurry is stabilized to high
mal cements encompass a wide variety of wellbore con- temperatures, and the collapse pressure (usually
ditions and complex chemical processes. Many factors 3,000 psi or 20.7 MPa) is not exceeded.
■ Foamed cement, made from a stabilized base slurry,
must be considered to determine the optimum cement
composition for a particular situation. Nevertheless, can be used with confidence in most thermal wells. In
there are several basic points that the engineer should geothermal wells, in which corrosive fluids are pro-
remember when contemplating this problem. duced, the long-term stability of foamed cements has
■ When static temperatures exceed 230°F [110°C], 35%
not been proven.
■ If the cement will be exposed to temperatures
to 40% silica BWOC must be added to Portland
cements; otherwise, strength retrogression will occur. exceeding 750°F [400°C], Portland cement should
■ If saline geothermal brines are present, fine silica
not be used. High-alumina cement is suitable.
■ Silica is deleterious to the stability of high-alumina
flour (less than 15-μm particle size) should be added
to Portland cement as a stabilizer. Silica sand does cements at temperatures exceeding 572°F [300°C].
not reliably provide adequate protection. Crushed aluminosilicate firebrick or fly ash is suit-
■ If high concentrations of CO2 are present, using cal-
able.
■ During laboratory testing, accurate static and circu-
cium aluminosilicate or calcium phosphate cements
is recommended. If Portland cement is used, degra- lating temperatures must be used to obtain an opti-
dation can be inhibited by reducing the silica con- mal thickening time and compressive strength at the
centration to 20% BWOC. wellsite.
■ Most common cement extenders are compatible with
thermal cements; however, if the static temperature
exceeds 450°F [232°C], fly ash should not be used in
Portland or Class J cement systems. Bentonite, per-
lite, and diatomaceous earth are suitable.
Table 11-1. Important Properties of Dry Additives with Respect to Cementing Logistics
Material Chemically Inert Chemically Active
Form Insoluble powder or finely cut material Soluble powder or finely cut material
Influence of accuracy in Concentration acts directly on system Materials may have secondary effects.
concentration on slurry quality density; no unexpected effect.
Handling The additives are blended with the dry The additives are normally dry-blended with the
cement in a special blender at the central cement. They are sometimes added to mix water
storage location up to several days before on location in an open horizontal tank, just prior
the pumping job. The blended material is to the job.
then transported to the wellsite. (If the
amount required is small and the material
easily scattered in the water, it is treated
as a soluble material.)
Storage Storage
Dry
cement (1)
Dry (3) Blend Surge
additive (3) tank
blending
HP steel
Dry
flow hoses
additives (2)
Water
Liquid Cement
Cement
additive slurry
mixer
mixing pumper
Liquid (2) Liquid
additives (2) additives (4)
Well casing
(1): Usually bulk, possibly U.S. (94-lbm) or metric (50-kg) sacks, or “big bags” (0.5 to 1.5 metric tons) topped with
(2): Manufacturer‘s packaging the cement
(3): Bulk (for very small and/or unplanned jobs, cement is sometimes stocked in paper sacks) head
(4): Bulk in special containers, except if mixing is done in an open tank
1.
2.
3.
4. 6.
1. Standard sacks or large bags
2. Bulk (notice the absence 5. Cement from the bulk
of surge tank on trailer) station sent to the
3. Sacks, large bags, or bulk rig location
4. Pneumatic silo 7.
5. Pneumatic loading bottle
6. Dry-additive blender 8.
7. Air-compressor plant
8. Horizontal tank trailer with
9.
merge tank
9. Twin vertical tank trailer
with merge tank
10. Supply boat, cementing 10.
barge, or vessel
Fig. 11-2. Delivery, storage, and distribution of cement and dry additives.
Air compressor
Surge
tank
Fill Vent
High-pressure Low-pressure
delivery delivery
Low-
pressure
Vent delivery
Fill
High-pressure delivery: Material transferred to storage tank or to mixing unit surge tank
Low-pressure delivery: Material air-blown to surge tank on rear of unit and dumped into cement mixer hopper
Back
Face
used. It is important to note that the pneumatic equip-
ment must be sufficiently powerful to blow heavy mate-
rials such as barite (specific gravity of 4.33) up to the
drilling rig tanks, a vertical distance of 130 to 200 ft [40 Porous
to 60 m]. The air compressors used for this task typically material
deliver 250 to 350 ft3/min [7,100 to 9,900 L/min] with a Air box
Valve Fill Vent
pressure rating of 28 to 44 psi [2 to 3 bar]. Sock
Air box
Delivery into
11-2.3 Wellsite storage of cement or blends Aeration and cement mixer
fluidization air hopper
As discussed above, pneumatic bulk trucks or trailers
transport neat or preblended dry cement to the wellsite Fig. 11-6. Atmospheric transportable bulk tank (typical piping
from the central storage and blending plant. Neat arrangement).
cement can also arrive directly from the cement mill.
The material is then transferred pneumatically to trans-
portable tanks that are either brought to the rig site for Safety valve
the cement job or are a permanent part of the drilling rig
equipment. Such tanks are similar to those used at cen-
tral storage locations, but their dimensions allow trans-
port on standard or specially designed (with a built-in
hydraulic laying/raising system) trailers. When empty,
the tanks must not exceed the weight limits specified by
various countries. A large variety of storage tanks for
road travel exists within two principal categories—
atmospheric and pressurized. Both are equipped with a
set of skids for proper installation on imperfectly leveled
ground and for easy winching onto trailers.
The atmospheric tank is always operated in a vertical
position. Air at low pressure (about 3 psi [0.2 bar]) is
blown into a gutter fixed to the slanted bottom of the 8 jets (1-in.)
2-in. pressurizing line
tank. The roof of the gutter is made of a porous material. 5-in. bleedoff/vent line
5-in. fill
The air passes through the porous partition and fluidizes 5-in. material delivery
Quick manhole
the cement blend. The cement blend glides along the 50-in. ID 6 jets (1-in.)
slanted bottom to a chute gate and then to the hopper of 2-in. air (letting) Bronze porous floor
a slurry mixing system. As illustrated in Fig. 11-6, atmos- 2-in. air (aeration)
pheric tanks are made in the shape of a parallelepiped.
Pressurized tanks use air at about 44-psi [3-bar] pres- Fig. 11-7. Pressurized bulk tank (typical piping arrangement).
sure and can operate horizontally or vertically.
Figure 11-7 is a schematic diagram of a typical unit. As
shown in Fig. 11-8, the vertical tanks are generally cylin- mixer. For versatility, some vertical pressurized tanks
droconical in shape, while horizontal models are more are also equipped to release the cement directly to a
complex. In the first stage, pressure-reduced air is blown hopper at atmospheric pressure (Fig. 11-9).
from the bottom through the mass of cement for aera- The bulk trailers are sometimes used for additional
tion and fluidization. Then air at 44 psi [3 bar] is storage. Indeed, they can serve all storage needs on the
injected into the tank, and the cement flows out through rig site, provided they are equipped with their own surge
a discharge line to a surge tank, which feeds the cement tanks, described later.
Bulk cement
(or blend)
Valve‡
Porous
Enlarged material Annular
to show air box
detail Valve† Valve†
Sock
Aeration and
fluidization air
Delivery to mixer hopper
Open in atmospheric mode, closed in high-pressure mode
†
Level scale
11-2.5.1 Liquid additive metering system with
metering tanks
All liquid-additive metering systems consist of two prin-
cipal parts—a storage and transfer unit and a metering
unit.
The storage/transfer unit generally includes four stor-
age tanks of various capacities (usually between 6.2 and
Drain
Open
25 bbl [1,000 and 4,000 L]). The storage and transfer
pipe Closed unit allows the independent metering of additives
according to the requirements of a particular job. This is
Mixing water
(to the mixing
convenient because well cement slurries typically con-
pump) tain two or three additives.
Each storage tank is equipped with its own air-oper-
Fig. 11-10. Displacement tank system. ated diaphragm pump and agitation system (recircula-
tion, as illustrated in Fig. 11-11, or air-operated stirrer)
to avoid segregation of the additive components.
For precise placement of the slurry in the wellbore, Therefore the operation of the unit requires a source of
the volume of the displacement fluid must be accurately clean and dry air at 120 to 145 psi [8 to 10 bar]. The con-
measured. After the cement slurry has passed through figuration of the unit varies depending on whether it is
the mixing system, the displacement fluid usually passes designed for use on land (skid or trailer-mounted) or off-
through the displacement tanks for volume measure- shore (containerized).
ment and is pumped by the cementing unit instead of The metering unit generally consists of a set of three
the rig’s mud pumps. (Fig. 11-11) or four 25-gal or 10-L tanks, with visible level
scales. To prepare a batch (10 bbl or 20 bbl [1.6 m3 or
3.18 m3] according to the displacement tank), the
11-2.5 Liquid additive storage and mixing
proper amounts of the selected additives are introduced
The simplest method of mixing liquid additives (and dry into the metering tanks. The additives are then released
additives at less than 3% by weight of cement) with into one of the two displacement-tank sections that is
water consists of pouring the required amount of each being filled with water. Finally, the mixture is agitated to
additive into a tank of water. One should measure the obtain a homogeneous solution. The same operation is
additives and water accurately to obtain the correct con- repeated for the following batch in the other half of the
centration; the preparation of a slight excess of solution displacement tank and so on. The repetitions of the
is also advisable. The mixing can be achieved with a operation may be automatically or semiautomatically
paddle mixer, circulation pump, jetting system, or a com- controlled.
bination of these.
The premix method has several disadvantages.
Premixing requires an extra tank, which must be clean
and sufficiently large. Extra tanks are not always avail-
able, and sufficient space to accommodate them may not
Water
Additive
storage
tank(s)
Air-powered
diaphragm
pump(s)
Mixing water
Vent
Dust separator 11-2.7.1 Conventional jet mixer
Air-forced The conventional jet mixer consists of a hopper, a mixing
cement bowl, a discharge gooseneck, and a slurry tub. The max-
Windows imum slurry-generating capacity of the conventional jet
(for watching mixer, evaluated in rate of dry material, is slightly higher
Air-jetting cement level) than 2,200 lbm/min (1 SI ton/min). Figure 11-14 shows a
system
(not shown) configuration for sacked cement, and a system for pneu-
matically delivered cement is illustrated in Fig. 11-15.
The cement is delivered to the hopper. The water is
Valve injected into the bowl through jets for mixing with the
cement and into the gooseneck for adjusting the slurry
Sock density. The jets are chosen according to the operating
pressure, slurry fabrication rate, and type of dry materi-
Delivery als. The movement of cement down through the hopper
(into mixer
hopper) is assisted by the high-pressure flow of water through the
jets. The resulting pressure drop pulls the dry cement
into the stream of water. To reinforce this effect, the
Fig. 11-13. Surge tank. gooseneck can be given a venturi tube profile. Further
along at the gooseneck, turbulent flow mixes the cement
particles with the water, and the result is a cement
slurry.
Mixing water
(from the mixing Slurry (to the
manifold) displacement
pump(s)
Additional
Knife water
Sack
Suction
Gooseneck pipe
Hopper Dry Grating
cement
Two or three jets
y
Slurr
Cutting Bowl
table
Additional
water
Butterfly valve
Suction
Sock pipe
Dry cement Gooseneck
Hopper Grating
Two or three jets
Bowl
The slurry density is adjusted by using the bypass ■ The slurry density is adjusted by operating the sliding
system to change the water-to-cement ratio. As the gate.
bypass is opened, the suction effect decreases and ■ The slurry is removed from the slurry tub by a recir-
reduces the amount of cement drawn out of the hopper. culation jet, fed by a centrifugal pump. The centrifu-
At the same time, the water bypassing the jets enters the gal pump force feeds the displacement pumps and
slurry. The combined effect is a decrease in slurry den- recirculates some slurry through the mixing system.
sity. Conversely, if the bypass is closed, the density ■ Water is always injected ahead of the recirculation
increases.
jet.
The conventional jet mixer can be operated at low
(175 to 200 psi [12 to 14 bar]) or high (880 to 1,180 psi Recirculation through the mixer heart and the tub
[60 to 80 bar]) water pressure. In the first case, the mix- improves the homogeneity and rheology of the slurry.
water pump is a centrifugal pump. In the latter case, it Adjustment of the slurry density is also easier.
is a reciprocating pump, usually identical (except per-
haps in plunger size) to the displacement pump. The
“double high-pressure pump cementing units,” which 11-2.7.3 Recirculation mixer without conventional
are the most widely used throughout the world, are jets
equipped to mix at either low or high pressure. The low- Available equipment includes a variety of mixers without
pressure method is preferred for two main reasons. Less conventional jets (Fig. 11-17). The maximum capacity of
horsepower is required and, because both high-pressure most mixers, evaluated by rate of dry material, is close to
pumps are available to displace the slurry, higher mixing 4,400 lbm/min [2 SI tons/min]. They all consist of the fol-
and displacement rates are possible. With the high-pres- lowing.
sure method, the jets and the bowl-and-gooseneck ■ A sophisticated metering system to mix cement with
assembly are less apt to become plugged with dirty water and a device to mix the resulting slurry with
mixing water or poor-quality cement. previously mixed slurry from the mixing tub
■ A centrifugal pump or similar device (located at the
11-2.7.2 Recirculation jet mixer bottom of the tub) to improve the initial mixing by
The maximum capacity of the recirculation jet mixer shearing, ensure recirculation through the mixer, and
(Fig. 11-16) is slightly more than 4,400 lbm/min feed pressurized slurry to the downhole pump
(2 SI ton/min). The recirculation jet mixer differs from ■ A mixing tub that can be divided into two sections,
the conventional type in several ways. each of which can be equipped with a stirrer to
■ A remotely controlled sliding gate is present between
improve mixing, allowing a film-like flow over the
the hopper and the bowl. common partition that assists the release of
entrapped air
Mixing water
(from the mixing
manifold) Slurry [to
Centrifugal the displacement
pump pumps(s)]
Dry cement
Suction
Remotely controlled pipe
sliding gate Grating
Tub
Jet(s) recirculating
line
Mixer
Slurry tub recirculating
line
Bulk cement
Cement
mixing valve
Mixing water
Water metering
valve (annular) Slurry
Recirculation
line
Centrifugal
pump
Cement
From mixing or blend
manifold Surge
Mixing tank
water
Densitometer
Slurry Slurry
To downhole To downhole
pump Centrifugal Centrifugal pump
Pump No. 1 Pump No. 2
Drain Drain
Available Available
Fig. 11-19. Twin-tank mixing unit (with recirculation jet mixer).
Surge
can
Tub level
Water
flowmeter
Slurry
flowmeter
Mixing
tub
Mixer Mixing water
To triplex
pump
Fig. 11-24. Schematic diagram of solids fraction monitoring equipment (Vigneaux et al., 2003). Reprinted with
permission of SPE.
field. In the near future, automatic cement mixing will Articulated (loop) section
undoubtedly become a routine procedure.
Three Two
11-2.10 Steel flowhoses and cement head swivels swivels
A “cement head” (Section 11-5.14) is screwed into the
top casing collar or landing joint, depending on the type
of cement job. The discharge side of the downhole pump Half union Half union
and the cement head are connected by a series of artic- (male) (female)
ulated or straight sections of high-pressure steel pipe,
also known as “treating iron” (Fig. 11-26).
Straight section
2
From rig
4 storage
Slurry
Water from to well
rig storage 9
1 5
10
7 8
1. Centrifugal water supply pump 6. Mixing water manifold
2. Water distributor 7. Cement mixer (conventional jet mixer shown)
3. Additive distributor 8. Slurry tub
4. Displacement tank system 9. Centrifugal pressurizing pump
5. Mixing water pump (centrifugal—low-pressure 10. Reciprocating displacement (downhole) pump(s)
mixing; reciprocating—high-pressure mixing)
Fig. 11-27. Mixing and pumping equipment on rig site (typical setup).
Cement
Recirculation
jet mixer
(not shown)
Pressurizing
pump Slurry tub Low-pressure Two high-
mixing pump pressure pumps
The safety requirements with which the equipment 1. Special water-cooled manifold rated to cool exhaust
should comply depend upon the location and are espe- gas to 200°C (392°F) maximum, and with a surface
cially dependent upon possible sources of flammable or temperature not exceeding 200°C at any point.
explosive gases. Whenever the unit can be placed more 2. Oversized radiator.
than 98 ft [30 m] away from the well (as on most land rig
3. Inlet air combustion, slam-shut valve.
sites) there are no special requirements. Standard
equipment can often be used without modification. This 4. Inlet air flame trap.
distance condition is often difficult to satisfy on offshore 5. Exhaust gas spark arrestor, DNV type approved.
rigs, where every compartment or deck location is clas- 6. Overspeed valve, which closes the engine blower
sified according to the potential risk of explosion or fire. flapper valve when speed exceeds the normal maxi-
The classification is made by official regulatory bodies mum by 10%.
according to standards that may vary slightly from one
7. High-water-coolant temperature valve, which shuts
country to another; however, operators usually adhere to
down the engine when water temperature exceeds
the most stringent regulations.
95°C (204°F). Fuel rack actuated.
For example, the following is a summary of the Det
Norske Veritas (DNV) requirements for diesel engines to 8. Low-water-coolant level valve, which shuts down the
be located in a hazardous area, classified as “Zone 2,” in engine.
which an explosive gas mixture may exist for a short time 9. High-exhaust-gas temperature valve, which shuts
only under abnormal conditions. Diesel engines are ban- down the engine when the gas temperature exceeds
ished from Zones 0 and 1, which are more sensitive 200°C.
areas. The DNV is the Norwegian certification body, and 10. Special control panel.
its standards serve as a reference in the North Sea.
■ Torque and drag reduction tools 11-5.1 Guide shoes and float shoes
An entire textbook could be written about these tools; Guide shoes and float shoes are tapered, commonly
for this textbook, the discussion will be limited to the bullet-nosed devices that are installed at the bottom of
most common or basic types, with the emphasis placed the casing string. They guide the casing toward the
on application, principles of operation, and basic design center of the hole to minimize hitting rock ledges or
characteristics. washouts as the casing is run into the well. The outer
portions of these shoes are usually made from steel, gen-
erally matching the casing in size and threads. The
Liner hanger
Casing packer
Spring-bow centralizer
Stage tool
Centralizer sub
Semirigid centralizer
Reamer shoe
Fig. 11-33. Typical application of casing hardware for a primary cement job (drawing courtesy of
Weatherford International).
Fig. 11-35. Cement nose guide shoe with down jet option.
Texas pattern
Air-filled Semirigid
casing centralizer
Mud
Fig. 11-40. Typical casing flotation system (drawing courtesy of Davis-Lynch, Inc.)
Seals
Circulating ports
Bottom sleeve
1. Bottom portion of casing is run dry 2. Casing pressure is increased until 3. Bottom cementing plug is launched
(not filled with fluid), with flotation collar the opening sleeve shifts down to ahead of cement. After landing on the
installed at desired depth. Casing above permit fluid and air to swap. After a bottom sleeve, it pushes both sleeves
the collar is filled with drilling fluid as fluid stabilization period, the casing ahead of the cement to the float collar
casing run continues to desired depth. is filled with drilling fluid. below.
Top plug
Casing
Casing
Bottom
plug
Bottom
plug
Float collar
Float collar
4. Bottom cementing plug and sleeves land and seal on 5. Top cementing plug seals and locks on bottom
the float collar. Bottom cementing plug ruptures, and cementing plug/collar assembly at the float collar.
cement is pumped through and out of the float equipment.
Fig. 11-41. Operating sequence of casing flotation system (drawing courtesy of Davis-Lynch, Inc.).
Disadvantages High erosion susceptibility Dependent on flow to close Low tolerance to some LCM (e.g., fibers)
Easily blocked Poor performance in inclined wellbores
Difficult to drill out High erosion susceptibility
Easily damaged