-
Copyright 2002, Society of Petroleum Engineers Inc. This paper
was prepared for presentation at the SPE Annual Technical
Conference and Exhibition held in San Antonio, Texas, 29 September2
October 2002. This paper was selected for presentation by an SPE
Program Committee following review of information contained in an
abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum
Engineers and are subject to correction by the author(s). The
material, as presented, does not necessarily reflect any position
of the Society of Petroleum Engineers, its officers, or members.
Papers presented at SPE meetings are subject to publication review
by Editorial Committees of the Society of Petroleum Engineers.
Electronic reproduction, distribution, or storage of any part of
this paper for commercial purposes without the written consent of
the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than
300 words; illustrations may not be copied. The abstract must
contain conspicuous acknowledgment of where and by whom the paper
was presented. Write Librarian, SPE, P.O. Box 833836, Richardson,
TX 75083-3836, U.S.A., fax 01-972-952-9435.
Abstract Cement pulsation is a novel technology for enhancing
zonal isolation by applying low frequency, hydraulic, pressure
pulses to the wellhead immediately after cementing. The treatment
maintains the slurry in a liquid state, which transmits hydrostatic
pressure downhole, and keeps the well overbalanced thus preventing
early gas flow after cementing.
The paper summarizes several stages in the development of cement
pulsation technology including comparison to other methods,
physical principles, process analysis, mathematical modeling,
computer-aided design, laboratory testing, and field
performance.
The paper supports published information on cement pulsation
with data from research and field studies that was instrumental in
developing the technology. Emphasis has been given to the analysis
of the pulsation process, description of design model and software,
and an updated account of field applications.
Described is the MS Windows software for pulsation design. Two
examples demonstrate the computer-aided design. The examples show
that the software could be used to find the pulse size and
treatment duration for a constant-pressure treatment.
Alternatively, a variable-pressure treatment with controlled
treatment depth could be designed.
Data is presented from pulsation of over 80 wells in drilling
areas notorious for early gas migration after cementing. Field
applications of the technology in 80 wells provided significant
evidence of the success of cement pulsation in preventing early gas
leakeage in cemented wells.
Introduction Top Cement Pulsation In 1982, a landmark field
experiment performed by Exxon revealed hydrostatic pressure loss in
the annuli after primary cementing in wells1. Since then,
hydrostatic pressure loss after cement placement has been
considered a primary reason for gas migration outside wells. As the
annular cement still in liquid state - loses hydrostatic pressure,
the well becomes under-balanced and formation gas invades the
slurry and finds its way upwards resulting in the loss of wells
integrity.
Cement slurry vibration using a low-frequency cyclic pulsation
is used by the construction industry for improving quality of
cement in terms of better compaction, compressive strength, and
fill-up. (Cement gelation or transmission of hydrostatic pressure
is not a concern in these applications.)
In the oil industry, the idea of keeping cement slurry in motion
after placement has been postulated a promising method for
prolonging slurry fluidity in order to sustain hydrostatic pressure
and prevent entry of gas into the wells annulus. The idea was based
upon experimental observations that cement slurries in continuous
motion remained liquidous for a prolonged period of time2,3.
Manipulating the casing string would move the cement slurry.
Thus, early concepts considered keeping cement slurry in motion
through casing rotation or reciprocation4,5,6. The motion should
improve displacement of drilling mud and placement of cement slurry
in the annulus.
The use of forced casing vibrations for gas flow control has
become subject of several inventions in the 80's and
90's7,8,9,10,11,12. For example, enhanced filling of annulus with
cement slurry without rotating or reciprocating the casing" was
considered the main advantage of the first casing vibration method
with mechanical vibrator placed at the bottom of the casing
string7. All these methods have been already experimentally studied
and patented. However, none of them have been used commercially
because of difficulty involved in manipulating the entire casing
string. Apparently, heavy equipment and installatioin needed to
vibrate a long and heavy string of casing makes these methods not
feasible, even onshore.
In 1995, Texaco patented a technique based on pulsation of the
cement top13,14. In this method, low frequency and small-amplitude
pressure pulses are applied at the top of the cement by cyclic
pumping of water or air to the wellhead. The
SPE 77752
Cement Pulsation Treatment in Wells Andrew K. Wojtanowicz, SPE,
John Rogers Smith, SPE, Djuro Novakovic, SPE/Louisiana State
University, V. S. Chimmalgi, SPE/ONGC, Ken R. Newman, SPE/Coiled
Tubing Engineering Services, Dale Dusterhoft, SPE/Trican Services,
Brian Gahan, SPE/Gas Technology Institute
-
2 WOJTANOWICZ, SMITH, NOVAKOVIC, CHIMMALGI, NEWMAN, DUSTERHOFT,
& GAHAN SPE 77752
treatment continues for sufficiently long time to keep cement in
liquid state, reduce transition time, and maintain hydrostatic
pressure overbalance.
Texaco field-tested a number of shallow (up to 4,700 feet) wells
in the Concho (Queen) field of the Permian basin, Texas. The tests
demonstrated that pulses could be transmitted through the slurry in
the lab and that the bond logs of pulsed wells were superior to
those that were not pulsed.
The Coiled Tubing Engineering Services, and the Louisiana State
University have jointly further developed the cement pusation
technology in a project sponsored by the Gas Technology Institute.
Field testing of instrumented wells (with downhole pressure gauges)
demonstrated that annular pulses could be transmitted to a
significant depth in excess of 9,000 ft and that hydrostatic
pressure in the annulus was maintained by pulsing the slurry15,16.
Full-scale laboratory pulsation experiments with a thixotropic
slurry in an LSU well showed how small pressure pulses would
progressively break gel structure and deliver pressure to the wells
bottom17,18. They also revealed that pulsation should have an
additional advantage versus application of a constant pressure18.
Another laboratory study showed that pulsation did not reduce final
compressive strength or shear bond of cement19.
Development and commercialization of the technology required a
method for designing the treatment. Mathematical modeling,
performed at LSU, provided theoretical basis for the treatment
design and diagnostic analysis methods and software3,17,20,21,22.
Industrial use of the technology has been carried out by Trican
Well Services Ltd. and Husky Energy in three oilfields of Eastern
Alberta, Canada23,24.
The objective of this paper is to support published information
on cement pulsation technology with data from research and field
studies that was instrumental in the technology development.
Emphasis has been given to the analysis of the pulsation process,
description of design model and software, and updated account of
field performnce.
Cement Pulsation Process and Equipment After cement placement,
the well annulus is intermittently pressurized-depressurized by
cyclically pumping water from the cement pulsation unit to the
wellhead. A portable cement pulsation unit consists of an air
compressor, water tank, hoses to connect to the well,
instrumentation and a recording system. Pulses are applied to the
annulus by water that is pressurized by the air compressor. After
charging the well, the water is bled back to the tank. The system
schematic is shown in Fig.1.
An air compressor continuously pressurizes an air tank. To
pressurize the annulus, the control system opens a valve between
the air tank and a water tank. The air pressure forces the water
into and pressurizes the casing annulus. To release the pressure,
the control system closes the pressurization valve and opens the
exhaust valve. As the pressure is released, water returns from the
casing annulus to the water tank. Once the pressure is fully
released, water is added to the water tank if needed, to keep the
water tank full.
The volume of water displaced to the well for each pulse is
determined by measuring the water level in the tank. From this
measurement a compressible volume is derived using a data
-smoothing algorithm with corrections for water loss in the well
and compressibility of surface installation.25. As the cement sets,
the compressible volume of the casing annulus should decrease as
shown in Fig. 2.
The pulses are quite slow, with built in delays. The pressure is
applied and held for up to 10-25 seconds (design parameter). After
pressure is released, there is a dormant period of up to 10- 25
seconds (design parameter). The pulsation frequency is low, of the
order of 1-2 cycle/minute (design parameter).
Recorded parameters of the pulsation process are shown in Fig.
3. Each cycle includes three periods, pre-pressurization,
pressurization, exhaust. During the pre-pressurization period, the
air-tank and water-tank are not in communication with each other.
The water that goes into the annulus during the previous cycle will
continue to come back and will result in water level increase. At
this time the annulus pressure can be in the range of 2 to 5 psi.
The compressor during this period is used to compress only the air
tank. The water tank pressure or annulus pressure will be very low
during this period.
When the pressurization is started, the use of pressurized air
to compress the water will reduce pressure in the air tank and
increase the pressure of water in the water tank. Once the air
pressure and the water pressure reach equilibrium, both the
pressures will continue to increase together, but at a lower rate
as some of the water is getting pumped into the annulus. During
this time, the pressure in the water tank will force the water out
of the tank into the annulus. The water level will be the lowest at
the maximum annulus or water-tank pressure.
When the air tank is cut off from the water tank and water
pressure is bled off during the exhaust period, the water pressure
will fall suddenly allowing the water that went into the annulus to
come back to the water tank. This will make the water level rise.
The compressor will continue to increase the air pressure in the
air tank during this period.
Physical Mechanism of Cement Pulsation The first question asked
about this technique is typically about energy efficiency. Unlike
other vibration techniques, this method uses small input energy
that is very efficiently transmitted over a long distance despite
the opposing friction. Thus, frictional losses must be small.
Efficient transmission of small top pressure pulses over several
thousand feet down the annular column of Non-Newtonian fluids with
yield stress could only be efficient if the column yields only at
the walls while the bulk fluid (a plug) remains not sheared. From
the analysis of flow of Bingham fluid in the annulus it has become
clear that, for the plug flow, energy required to reciprocate the
slurry is much smaller than that needed to shear the entire bulk
slurry.
The analysis required revisiting the theory of Bingham fluid and
deriving exact formulas for plug flow17,24. Figure 4 demonstrates
the concept of a minimum velocity needed to shear the bulk slurry
(reduce plug size to zero). It has been
-
SPE 77752 CEMENT PULSATION TREATMENT IN WELLS 3
shown that for typical pulstion parameters and annular sizes the
slurry motion is well within the plug flow regime24. Moreover, the
exact model of plug flow gives small values of shearing rates and
pressure losses (energy loss) at velocities representing cement
pulsation. The shearing rate relation is shown in Fig. 5, and the
pressure loss in Ref. 17.
As the pulsed slurry moves only as a plug, its reciprocating
motion can be simplified as an equivalent slow continuous motion in
the plug flow regime. Thus, the relationship between average
velocity and displacement amplitude becomes,
fzyzv *)(21)( = (1)
Displacement amplitude, y(z), is distributed along the slurry
column. Cement in the upper annulus undergoes greater displacement
than the deeper cement - the displacement amplitude reduces with
increasing depth so does the friction opposing the displacement as
the velocity is also reduced with depth.
After a few modeling attempts aimed at pressure wave propagation
and other effects of pressure transient, the modeling focused on a
pseudo dynamic concept where the velocity of the fluid and the
deformation of the annular walls is considered, while the transient
effects are neglected. The simplification is based on evaluation of
the pressure pulse propagation velocity in the annular system for
different annular sizes24. It was found that the velocity of
pressure wave in the annulus would range between 2500 ft/second to
4200 ft/sec. As the typical pressure pulse duration used in cement
pulsation is of the order of the 10s of seconds and the velocity of
pulse application is low, the pressure transients are
negligible.
Keeping slurry in motion reduces static gel strength (SGS)
development and delays the transition from liquid to solid. Several
experiments have shown that the change of SGS development is
practically independent from shearing rate.(3) Thus, for the
design, one could assume that as long as the slurry is sheared at
the wall, its gelation could be represented by a single altered SGS
plot disregarding the shearing rate value. This also means that the
pulsation treatment is effective when pressure value at depth
exceeds the time-dependent value of the altered SGS of the slurry.
The condition would define the depth of (effective) treatment.
Two sections could be visualized in the pulse-treated cement
column. The upper, and usually very long, section is where the
shear stress at the wall is larger than yield stress of the slurry
and the slurry is in motion. In the bottom, typically short,
section, the transmitted pressure pulse is smaller than yield
stress so the slurry is motionless and not sheared. In this work we
assume that the latter section column is negligibly small and all
the pressure is expended when the displacement amplitude becomes
zero. Hence in our analysis the terms treated depth and the depth
of pressure transmission are equal.
During pulsation, the cement slurry at depth is sheared at the
walls as it reciprocates upwards and downwards. As the annulus is a
closed system, the slurry movement in the annulus
is caused by elastic deformation of the fluid and the annular
walls. Thus, the displacement amplitude is caused by the pressure
at a given depth and is controlled by the compressibility of the
annulus below that depth. The annular compressibility represents
isothermal compressibility of the slurry coupled with the elastic
properties of the cased hole and a stratified open hole built of
several layers of rocks having different properties and thickness.
The annular system compressibility model has been derived and
presented elsewhere17,26.
Designing pulsation treatment for a well involves determination
of parameters such as the pressure pulse amplitude, pulse cycle
duration, and depth of treatment as functions of time.
Interestingly, the three parameters are somewhat dependent on each
other as well as on the properties of cement, mud, and rock, and on
the annular geometry of the well. Hence, mathematical modeling of
pressure pulse transmission (attenuation), and displacement
amplitude distribution becomes a basis for the design.
Mathematical Model of Cement Pulsation Process Derivation of the
mathematical model has been based on the following assumptions:
Reciprocating motion of slurry is represented by
equivalent continuous motion with average velocity given by Eq.
(1);
The annular fluid motion follows Bingham plastic model in plug
flow;
System compressibility applies and the annular system deforms
elastically;
Pressure pulse duration is sufficiently long so inertial effects
can be neglected;
Time dependent properties of stagnant and sheared slurry (yield
stress and plastic viscosity) are known from pre-job testing;
Duration of time lapse between the two consecutive pulses is
sufficently long so that the stress from the previous pulse fully
diminishes; i.e. displacement amplitude is not affected by residual
stresses;
There is an active mechanism of stress relaxation in the annular
fluid column fluid loss to the rock;
The applied top pressure is attenuated by the total distributed
friction due to slurry movement; i.e frictional pressure loss in
plug flow controls pressure transmission downhole;
Displacement amplitude is distributed and controlled by
compressibility and pressure distributions.
The top pressure pulse, p0, transmission formula is:
dzGvKdpz
po
)(0
0
+= (2) where:
-
4 WOJTANOWICZ, SMITH, NOVAKOVIC, CHIMMALGI, NEWMAN, DUSTERHOFT,
& GAHAN SPE 77752
12
12
212
ln0614.0936.0
)(200
)(1000
ddvC
ddYPC
G
ddPVC
K
f
f
f
+=
==
(3)
Displacement amplitude is described by the linear differential
equation:
)]1)[exp(1)(5.0)exp( cGzcpzczKfycpdzdy
oo +=+(4)
which gives distributed displacement formula:
+
+=
3)exp(
3)exp())exp(1()(
32
32
0azzaz
aZZaZcpzy ppp
(5) where:
)exp(25.0 0cpcKfa = (6) From the model, the depth of treatment,
Zp, and the top displacement amplitude, Y0, are calculated from the
equation:
+=
3)exp())exp(1(
32
0p
ppo
aZZaZcpY (7)
Finally, the bottomhole pressure at any time is computed as:
))(4())(4()(1212
pp ZddYSgZZ
ddSGSgZp += (8)
The mathematical model was validated in full-scale pulsation
experiments at the LSU Well facility, and by matching data recorded
during cement pulsation in two wells in Texas17,26. We also used
this model in a sensitivity study to evaluate relative effect of
the parameters involved in the pulsation process. The study
revealed that: Large well annuli improve pulse transmission
significantly; There is almost linear increase of treatment
depth with the
size of top pressure pulse; Low-frequency pulses (f < 0.1)
would significantly
increase treatment depth; Annular system with large
compressibility would reduce
treatment depth.
Cement Pulsation Design Software Cement Pulsation Design
software is a MS Excel based application integrating both
spreadsheet calculations to handle the local mathematics and VB
Macros to handle global task assignment and property distribution
within the well of interest. The Main spreadsheet is shown in Fig.
6. The software consists of several spreadsheets, designed to
separate several aspects of data. (User can navigate between
spreadsheets either by clicking the button containing particular
aspect of data input or by using tabs in the bottom of the screen.)
The input spreadsheet names and button titles for corresponding
sheets are: ControlPanel, OpenHole, CasedHole, and Fluids. The
output data is saved in the spreadsheet Results.
The ControlPanel spreadsheet is used to enable/disable the check
for necessary Excel Add-Ins (Solver and Analysis ToolPak) or
support software provided by Microsoft with original installation
CD. Also using the ControlPanel, the user may request a customized
listing of time-related properties at specific depths of interest.
If the listing were omitted, the software would output the treated
depth, top displacement and top pressure pulse for each
timestep.
The OpenHole spreadsheet takes input data for up to 20 different
rock strata composing the openhole section of the well annulus. For
each strata the input is similar to that for the CasedHole section
and includes: bottom depth of the strata, size of vertical
gridblocks, sizes (open hole diameter, inner and outer casing
diameter), Poisson's ratio for rock and steel, Young's modulus for
rock and steel, and compressibility of the fluid opposite the
strata.
Values of Youngs modulus and Poissons ratio for the openhole
strata can be estimated from empirical correlations. Poisson ratio
values for sedimentary rocks vary from 0.2 for a hard and fragile
rock to 0.5 for most soft rocks. Correlations of Poison ratio vs.
confining/overburden pressure have been developed for specific
areas. For the Gulf coast area, the Eaton correlation gives the
Poison ratio vs. depth directly. For other areas, the Poissons
ratio can be calculated using Eaton correlation for Gulf coast area
with variable overburden pressure. The procedure for finding the
overburden pressure is described elsewhere27. Once the overburden
pressure is known, Youngs modulus can be computed from the
empirical formula,
naa pppKE )/(0= (9)
where, K0 = rock dependent empirical modulus number n = rock
dependent exponent pa = atmospheric pressure p = overburden
pressure Values for K0 and n for elastic rocks are given in Table
3.1
in Ref. 28. The remaing input data are saved in the Fluids
spreadheet.
The spreadsheet accepts two types of time-dependent data,
annular fluids properties (drilling mud, tail, and lead slurry),
and pulsation data (pressure pulse size, and cycle time).
The Results spreadsheet returns two types of output data: time-
dependent and depth-dependent. The time dependent results are time,
minutes; treatment depth, ft; top displacement, gal; and top
pressure, psi. The depth-dependent results are controlled by a
user-provided listing in the ControlPanel and they include a header
containing timestep and pulse cycle; depth, ft; pulse pressure,
psi; and, displacement, gal.
-
SPE 77752 CEMENT PULSATION TREATMENT IN WELLS 5
Algorithm By definition, average fluid properties for the
annular system are computed from Eqs. (10) and (11). Equation (12)
gives the first approximation of depth of pressure pulse
transmission, which is a function only of the average yield point.
In the subsequent iterations the treatment depth is computed from
Eq. (13) derived from Eq. (2).
ip
iN
iavg PVZ
lPV )()(1=
= (10)
ip
iN
iavg YSZ
lYS )()(1=
= (11)
avgp YP
ddpZ )(300 120 = (12)
dzGvKppZ
)(0
0 += (13) The equations (3,5,6,10,11, and 13) are iterated until
the
depth of treatment, average Yield Point, and Plastic Viscosity
converge.
Compressibility of the total annular system (down to the cement
bottom) remains constant throughout the pulsation process3.
However, the compressibility of the treated annulus is not
constant. Most of the time, the treated annular system will have
both open hole and cased hole portions, with the openhole section
having larger compressibility than the cased hole section. (By
definition, the system compressibility is the volumetric averages
of the open hole compressibility and the cased hole
compressibility.) Thus, with continuing pulsation and increasing
gelation, the depth of pressure pulse propagation (treatment depth)
and the treated system compressibility will both reduce.
Consequently, the treated system compressibility values must be
updated after each series of iterations.
The software algorithm flowchart, presented in Fig. 7, can be
summarized as follows. For the previous value of treatment depth,
the average PV and YP (and system compressibility) is computed from
Eqs. (10) and (11). Then, the top displacement amplitude is
computed from Eq. (7). Since the correction factor Cf is velocity
dependent and we do not know the velocities of the cement column,
we assume unity value for Cf (simplified plug flow model) and
calculate the top displacement amplitude and velocity from Eq.
(1).
The annular column is divided into large number of small grids
to enable us to take into account the variation in the annular
geometry and the fluid properties. The displacement amplitude,
velocity, correction factor, Cf, and frictional pressure loss is
calculated across each grid together with cumulative pressure
attenuation described by the integral in Eq. (2). The calculation
for different grid blocks is continued till the transmitted
pressure pulse becomes zero. The corresponding depth becomes an
updated depth of treatment
and is substituted back to calculate once again the average
property of the fluid, the corresponding system compressibility and
top displacement. Once this operation is carried out for the
grid-wise pressure distribution, the spreadsheet automatically
updates displacement amplitude distributions. These iterations are
repeated until the depth used to calculate the top displacement and
the depth of pressure pulse transmission become equal. Normally,
the calculations converge within two to three iterations. The above
computations represent the calculation procedure for a single time
step.
For next time step, the fluid properties are updated and the
procedure of calculation is repeated. The top displacements and the
depth of treatment for each of the time step are saved and plotted
against time or number of pulses. Also, as the iterations are
carried out for a non-linear system of equations, the safest way to
do it is using MS Excel's Add-in Solver.
Examples of Computer-aided Design Two types of cement pulsation
treatment are demonstrated below, constant-pressure operations and
controlled-depth operations. The treatments were designed for the
same well in Texas shown in Fig. 8. (The subject well was actually
treated using the constant-pressure pulsation pattern.) The
properties of annular fluid used in the design were as follows:
Drilling mud:
Depth = 6,900 ft Density = 12.5 ppg YP = 10 lb/100 sq ft PV = 18
cp
Lead cement: Depth = 7,700 ft Density = 12.5 ppg YS = (
)200198.09681.8 + te PV = ( )200169.09917.9 + te Tail cement:
Density = 15 ppg YS = ( )200053.0674.70 + te PV = 65.1491124.210510
2336 +++ ttt Where t is elapsed time in minutes.
Predicted results of the constant-pulsation process are shown in
Figs. 9 and 10. Over the first 50 minutes, the top 300 feet of the
lead slurry column is pulsed followed by continuing pulsation of
the lead slurry until 2 hr and 10 min of the treatment. After that
time, only drilling mud column is pulsed; the treated depth becomes
constant and equal to 6,900 ft. Also, the top displacement
amplitude becomes constant and equal to 10 ft Fig. 10.
The controlled-depth pulsation was designed in two stages.
During the first 120 minutes pulsation is carried out with constant
pressure of 100 psi Fig 13. During that time, the depth of
treatment reduces from the initial 8,000 ft to 7,000 ft, i.e. to
the top 100 feet of the lead cement. To treat more cement for
longer time we plan to increase the treatment depth from 7,000 ft
to 7,100 ft and keep the depth constant until 3
-
6 WOJTANOWICZ, SMITH, NOVAKOVIC, CHIMMALGI, NEWMAN, DUSTERHOFT,
& GAHAN SPE 77752
hrs of pulsation as shown in Fig. 11. The increased and constant
treatment depth requires progressive enlargment of the top pressure
pulse from 100 psi to 230 psi Fig. 13. The increase of top pressure
compresses mostly the 6,900-foot long mud column so the top
displacement amplitude should rise. In fact, figure 12 clearly
demonstrates the increase of the amplitude from 11 to 25 feet.
Laboratory Testing of Pulsed Cement Slurries The proposed
testing protocol for cement pulsation is analogous to, but
different than, the pre-job tests performed for primary cementing
operations. This testing is not mandatory, but it is useful for
determining the feasibility of, and for making both simple and
comprehensive performance predictions for, a specific job. This is
particularly important for deep jobs, for those with a high
viscosity drilling fluid, or for diagnosing problems during a
job.
The most important properties are a pulsation-specific yield
point, conventional gel strength measurements, and the expected
pulsation time analogous to thickening time. Modified cement test
methods can also be used to demonstrate how pulsation controls the
development of cement gel strength. The protocol uses conventional
lab equipment to measure the required mud and cement properties.
The specific procedures for measuring all these properties, the
basis for such measurements, and an example application to an
instrumented field cement pulsation job were previously described
in detail by Smith et al29.
Tests of Mud Properties. The mud properties that must be
measured for use with cement pulsation are the plastic viscosity,
the yield point, and the gel strength versus time. Viscometer
speeds of 3, 6, and 100 rpm were selected to represent the range of
fluid velocities expected in the well annulus during pulsation. The
following definition of yield point was validated by previous
work18. The pulsation specific mud rheology parameters are: YP = 3
6 3( ) (14) PV YP= 3 100( ) (15) The mud gel strength is measured
using conventional definitions and procedures for drilling
fluids27.
Tests of Cement Properties. The cement properties that should be
measured are the same as for mud. However for cement, all of the
properties vary with time and prior shear history. The viscometer
measurements are made with a specially modified Fann viscometer.
Measurements are made at 3, 6, and 30 rpm with a standard F1 spring
and a 1.2276 cm radius bob that has been knurled to minimize cement
slippage. The speeds were selected to represent the range of
annular velocities expected in the cement column. YP k k k= 2 11 3
6 3. ( ( )) (16) PV YPk= 29 7 2 1130. ( ( / . )) (17) Gel Strength
k peak= 211 3. , (18)
The validity of these measurements was confirmed by comparison
of simple predictions of job feasibility and performance to
downhole pressure measurements reported in a previous paper29.
Measurement of the maximum cement pulsation treatment time is
performed by operating a MACS Analyzer at 8 rpm to simulate
pulsation. A consistency reading of 25 to 35 Bc at 8 rpm is
proposed as the limit which determines the maximum treatment time.
The actual static gel strength can also be measured when the
consistency reaches this level by stopping rotation and switching
the device to its static gel strength measuring mode.
The measurement of conventional static gel strength versus time
under normal conditions is also potentially useful. Comparing the
consistency versus time under simulated cement pulsation with
static gel strength versus time can give an approximate indication
of whether pulsation will suppress gel strength development for a
particular slurry.
Field Results The final phase of the research was to test the
pulsation theory in the field to determine if pulses would, in
fact, prevent gas flow into the annulus. These tests could confirm
that the ability of cement pulsation to maintain bottom hole
pressure demonstrated in previous instrumented field tests15,16
does supress flow after cementing. Therefore, cement pulsation has
been applied to a total of eighty wells in Canada in areas with
previous gas migration problems. A more complete description of the
test program is given by Dusterhoft et al24.
Wells in the Eastern Alberta area of Canada were chosen to test
the technique. This area was considered ideal since it contains a
number of wells that experience gas migration to surface, which is
easy to measure by monitoring surface casing vents. Another
advantage of this area is that it has a solid history of recorded
vent leaks to compare to.
A pulsation cementing project was undertaken with Husky Energy
in three fields: the Tangleflags, Wildmere and Abbey fields. All of
these fields have experienced various levels of gas migration in
the past and numerous techniques have been used in an attempt to
control the problems.
Tangleflags Area. A typical Tangleflags well is described in
Table 1. This area has had a history of moderate gas migration
problems with an average of 10.5% of the wells drilled experiencing
gas migration problems. A total of 24 wells were included in the
study: seventeen were pulsed and two were abandonment plugs that
were also pulsed. None of the wells pulsed experienced any
leaks.
Wildmere Area. A typical Wildmere well is also described in
Table 1. It has been more difficult to control gas migration in
Wildmere, which has had an average of 25% of the wells leaking.
Twenty wells were cemented in this area: four were plug jobs and 16
were production cement jobs. Of the 20 cemented, 18 had no leaks
while two wells leaked. Both wells that leaked experienced
equipment freezing problems during pulsation so the pulses were not
transmitted to the annulus.
-
SPE 77752 CEMENT PULSATION TREATMENT IN WELLS 7
The above test results were taken from the information available
from Huskys gas migration test database and may not be inclusive.
Not all wells drilled in the years specified may be reflected in
the numbers - some wells do not have tests on file. The
leaker/non-leaker status is based on the most recent test
conducted.
Abbey. Table 2 outlines a typical Abbey well. Abbey is in a
river valley and wells in this area have serious gas migration
problems. The wells here are shallower due to the lower surface
elevations. Gas migration is believed to originate from the Milk
River zone at 360 m and 3,800 kPa. This depth reduction, in
conjunction with higher pressure gas zones, has resulted in more
than 80% of the wells leaking in this field. Eight jobs have been
performed in this area. Six of the wells have not leaked and two
wells have leaked. One well is leaking through the vent and the
second is leaking through the ground around the surface casing.
Additional Field Testing. Since the initial Husky tests, an
additional 28 wells have been cemented in other serious gas
migration areas of Alberta. To date, the pulsation system has been
100% successful in preventing vent leaks in these wells. The
deepest wells pulsed to date have been 1,300 m. Four wells were
cemented to this depth in an area where 75% of the offset wells had
vent leaks. All four pulsed wells had no vent leaks.
One two-stage well was pulsed. The second stage was cemented
from 1,000 m to surface and was pulsed. The previous two offset
wells experienced surface vent leaks from a gas zone above 1,000 m.
The pulsed well had no vent leak. A summary of the additional wells
is contained in Table 3.
Conclusions The paper summarizes several stages in developing
cement pulsation technology, from comparison to other methods, to
physical principles, to process analysis, to mathematical modeling,
to computer-aided design, to laboratory testing, and, to field
performance. Several aspects of the technology have been supported
with new data leading to the following conclusions: 1. Keeping the
cement slurry in motion prolongs its liquidity
and ensures well pressure overbalance. The motion can be induced
with several techniques: casing/cement vibration, casing
rotation/reciprocation, or cement pulsation. The latter method is
the simplest, most convenient to use without disrupting rig
routines.
2. Low-frequency and small-amplitude pressure pulses require
small input energy but can be transmitted deep downhole with little
attenuation. The reason for high energy efficiency is the plug flow
motion of the slurry with small energy loss for shear friction
outside the friction-less plug.
3. The pulse cycle time is long and greater than the transit
time. The pulsation process involves a series of individual
applications of pressure to the top of the annulus. In each
application the pressure is held long enough so it could be
felt deep downhole. This principle not only defines the designed
duration of the pressure holding time, but it also allows
elimination of pressure transient effect for mathematical modeling
of this process.
4. The mathematical model describes effect of a single pressure
pulse at a time. (The rectangular pulse shape has been verified
with the monitoring system of the pulsation unit.) The model
relates the applied pressure to the length of the annular slurry
column in motion (treatment depth) for known dynamic (in-motion)
properties (PV, YS) of the slurry at that time. As YSdynamic <
SGSstatic; the hydrostatic pressure at depth is greater than that
for a static column at any time.
5. The design software simulates cement pulsation process by
recurrent pulse-wise applications of the mathematical model over
the entire treatment time involving hundreds of pulses. At each
pulse, the software searches for iterative solution to the system
of non-linear equations describing pressure/dispalcemnt
transmission along the heterogenous annular column comprising
several sections of different fluids and up to 20 layers of rocks
in the open hole.
6. As demonstrated in examples, the cement pulsation software
could be used to find the pulse size and treatment duration for a
constant-pressure treatment. Alternatively, a variable-pressure
treatment with controlled treatment depth could be designed.
7. The fluid property definitions proposed herein provide a
reasonable basis for predictions of cement pulsation feasibility
and performance.
8. Field application of the technology in 80 wells provided
statictically significant proof of cement pulsation performance in
preventing early gas leaking in cemented wells.
Nomenclature a = flow parameter defined by Eq. (6) c =
compressibility, 1/psi Cf = parameter defined by Eq. (3) d1 and d2
= annular diameters: inside and outside, respectively E = Youngs
modulus of rock, psi f = frequency, 1/sec G = flow parameter
defined by Eq. (3) K = flow parameter defined by Eq. (3) K0 = rock
modulus number, 103 in. l1, li...lN = lengths of different fluid
sections, ft N = number of fluid sections in annulus = viscometer
dial reading PV (t) = plastic viscosity, cp p0 = top pressure
pulse, psi p = frictional pressure loss t = time, min _v = average
velocity of pulsed slurry defined by Eq. (1) z = depth, ft Z =
depth of cement bottom, ft Zp = depth of pressure pulse travel
(treatment depth), ft
-
8 WOJTANOWICZ, SMITH, NOVAKOVIC, CHIMMALGI, NEWMAN, DUSTERHOFT,
& GAHAN SPE 77752
y(z)= displacement amplitude at depth, ft Y0 = top displacement
amplitude, ft YP (t) = yield point (mud: YP=YS), lbf/100sq.ft. YS
(t) = dynamic yield stress of pulsed cement, lbf/100sq.ft. SGS =
static gel strength, lbf/100sq.ft. = density, lb/gal g =
acceleration of gravity, lb-ft/sec2 References 1. Cooke C.E. Jr.,
Kluck M.P., and Medrano R.: "Field
Measurements of Annular Pressure and Temperature During Primary
Cementing", SPE Paper 11206, 1982.
2. Cooke, C.E., Gonzalez, O.J., and Broussard D.J.: Primary
Cementing Improvement by Casing Vibration During Cementing Casing
Time, SPE 14199, 1988.
3. Wojtanowicz, A.K., and Manowski, W.: "Pressure Pulsation of
Cement for Improved Well Integrity - Field Method and Theoretical
Model," Proc. 10th Int. Scientific & Technical Conference: New
Methods and Technologies in Petroleum Geology, Drilling and
Reservoir Engineering, Krakow, Poland, June 24-25, 1999, Vol. 2,
421-436.
4. Carter, G., and Slagle, K.: A Study of Completion Pratices to
Minimize Gas Communication, SPE 3164, Central Plains Regional
Meeting of the Society of Petroleum Engineers of AIME, Amarillo, TX
(Nov. 16 -17, 1970).
5. Carter, G., Cook, C., and Snelson, L.: Cementing Research in
Directional Gas Well Completions, SPE 4313, Second Annual European
Meeting of the Society of Petroleum Engineers of AIME, London,
England (Apr. 2-3, 1973)
6. Christian, W.W., Chatterji, J., and Ostroot G.: Gas Leakage
in Primary Cementing - A Field Study and Laboratory Investigation,
SPE 5517, 50th Annual Fall Meeting of the Society of Petroleum
Engineers of AIME, Dallas, TX (Sept. 28 - Oct. 1, 1975).
7. Solum, K.W. et al.: Method and Apparatus for Vibrating and
Cementing a Well Casing, U.S. Patent 3,557,875, Jan. 26, 1971.
8. Cooke, C.E. Jr.: Method for Preventing Annular Fluid Flow,
U.S. Patent 4,407,365, Oct. 4, 1983.
9. Keller, S.R.: Oscillatory Flow Method for Improved Well
Cementing, U.S. Patent 4,548,271, Oct. 22, 1985.
10. Bodine, A.G., and Gregory, Y.N.: Sonic Cementing, U.S.
Patent 4,640,360, Feb. 3, 1987.
11. Rankin, R.E., and Rankin, K.T.: Apparatus and Method for
Vibrating a Casing String During Cementing, U.S. Patent 5,152,342,
Oct. 6, 1992.
12. Winbow, G. A.: Method for Preventing Annular Fluid Flow
Using Tube Waves, U.S. Patent 5,361,837, Nov. 8, 1994.
13. Haberman J. P., Delestatius D. M., and Brace D.G.: Method
and Apparatus to Improve the Displacement of Drilling Fluid by
Cement Slurries During Primary and Remedial Cementing Operations,
to Improve Cement Bond Logs and to Reduce or Eliminate Gas
Migration Problems, US Patent 6,645,661, 1995.
14. Haberman J. P., and Wolhart, S.L.: Reciprocating Cement
Slurries After Placement by Applying Pressure Pulses in the
Annulus, SPE/IADC 37619, March 1997.
15. Newman, K., Wojtanowicz, A.K., and Gahan, B.C.: Improving
Gas Well Cement Jobs with Cement Pulsation, Gas Tips, Fall 2001,
pp. 29 33.
16. Newman, K., Wojtanowicz, A.K., and Gahan, B.C.: Cement
Pulsation Improves Gas Well Cementing, World Oil, July 2001, pp. 89
94.
17. Chimmalgi, V.S., and Wojtanowicz, A.K.: Design of Cement
Pulsation Treatment in Gas Wells Model and Field Validation, Paper
2002-240, Petroleum Societys Canadian Petroleum Conference 2002,
Calgary, Alberta, Canada, June 11-13, 2002.
18. Martin, J.N., Smith J.R., and Wojtanowicz, A.K,:
Experimental Assessment of Methods to Maintain Bottomhole Pressure
After Cement Placement, ETCE01-17133, ASME Engiuneering Technology
Conference on Energy, ETCE 2001, February 5-7, 2001, Houston,
TX.
19. Shear Bond/Compressive Strength testing, CSI Final Report on
Pulsation Project submitted to CTES, Houston, TX, 2001
20. Manowski, W.M., and Wojtanowicz, A.K.: Oilwell Cement
Pulsing to Maintain Hydrostatic Pressure: A Search for Design
Model, J. Energy Resource Technology-Transactions of the ASME, Vol
120, December 1998, pp 250-255.
21. Kunju, M.R., and Wojtanowicz, A.K.: Well Cementing Diagnosis
from Top Cement Pulsation Record, SPE 71387, SPE Annual Technical
Conference and Exhibition, New Orleans, LA, September 30 October 3,
2001.
22. Novakovic, D., Wojtanowicz, A.K, and Chimmalgi V.S.: Cement
Pulsation Design Software, LSU Final report submitted to GTI
(August 2001) 42.
23. Dusterhoft, D., and Wilson, G.: Field Study of the Use of
Cement Pulsation to Control Gas Migration, Paper 2001-01 presented
at the C ADE/CAODC Drilling Conference, Calgary, Alberta, Canada,
October 23-24, 2001.
24. Dusterhoft, D., Wilson, G., and Newman, K.: Field Study of
the Use of Cement Pulsation to Control Gas Migration, SPE 75689,
SPE Gas Technology Symposium, Calgary, Alberta, Canada, April
30-May 2, 2002.
25. Kunju, M.R.: Post-treatment Diagnosis of Cement Pulsation in
Wells, MS Thesis, Chapter 4, Louisiana State University (May 2001)
33.
26. Chimmalgi, V.S.: Design of Cement Top pulsation to Avoid Gas
Migration During Cementing, MS Thesis, Chapter 2, Louisiana State
University (December 2001) 37.
27. Bourgoyne, A. T. Jr et al.: Applied Drilling Engineering,
second edition, Society of Petroleum Engineers, Richardson, TX
(1991) 42, 502.
28. Gidley, J.L., Hoditch, S.A., Nierode, D.E., and Veatch, R.W.
Jr.: Recent Advances in Hydraulkic Fracturing, first printing,
Society of Petroleum Engineers, Richardson, TX (1989) 452.
29. Smith, J.R., Martin, J.N., Newman, K. R. and Gahan, B.C.:
Field Evaluation of Pre-Job Test Protocol for Cement Pulsation,
CIPC 2002, Calgary, AB, June 11-13, 2002.
-
SPE 77752 CEMENT PULSATION TREATMENT IN WELLS 9
Table 1 Typical Tangleflags and Wildmere Wells
Tangleflags Wildmere
Location: 51-26-W3M 48-6-W4M
TD: Approximately 600 m Approximately 700 m
Casing: 177.8 177.8
Hole Size: 222.3 222.3
Cement Tops Surface Surface
BHST: 25oC 25oC
BHCT: 25oC 25oC
Surface Casing Depth 133 m 133 m
Gas Producing Zones Up and down the hole Up and down the
hole
Table 2 Typical Abbey Well
Location: 22-17-W3M
Surface 244.5 m casing at 70 m
Intermediate: 177.8 mm casing at 130 m
222.3 mm hole size
BHST = 90oC; BHCT = 23oC
Potential gas zones along interval
Production: 114.3 casing to 500 m
158.8 mm hole size
BHST = 26oC; BHCT = 23oC
Potential Gas Zones Milk River at 360 m BHP of 3,800 kPa
Table 3 Additional Pulsed Wells
Area No. of Wells Avg. Depth (m) Estimate of % Vent
Leaks Prior to Pulsation
% Vent Leaks After Pulsation
Lloydminster 23 800 20% 0%
Red Deer 1 1000 100% on 2 offsets 0%
Whitecourt 4 1300 75% 0%
-
10 WOJTANOWICZ, SMITH, NOVAKOVIC, CHIMMALGI, NEWMAN, DUSTERHOFT,
& GAHAN SPE 77752
200 gal.200 psi V.P.
Water to Well Annulus
Water Input Water Tank
Air Tank
Air control valveAir Input
200 gal200 psi V.P.
Figure 1- Cement pulsation unit flowpath
2
3
4
5
6
7
8
9
10
11
12
0 30 60 90 120 150 180 210 240Time (Minutes)
Vol
ume
(Gal
lons
)
Figure 2 Change of compressible volume during cement
pulsation
-
SPE 77752 CEMENT PULSATION TREATMENT IN WELLS 11
0
20
40
60
80
100
120
140
160
180
64 74 84 94 104 114 124 134Time (Sec)
Pre
ssur
e(ps
i) / v
olum
e (g
al)
Annulus/ Hose Pressure Air Tank Pressure Water Level
Pre-PressurePressurize
Exhaust
Pre-Pressure
Cycle 2Cycle 1
Figure 3 Recorded parameters of cement pulsation cycle
Figure 4 Plug size reduction with increasing flow velocity
12.25X9.625 Annulus; YP = 40 lb/100sq.ft.; PV=83 cp
0
20
40
60
80
100
120
0.03 0.05 0.10 0.25 0.30 0.40 0.50 0.75 1.00 1.25 1.50 1.75 2.00
3.00 4.00 5.00 6.00 7.00 8.00 Velocity, ft/sec
Plu
g S
ize/
Ann
ulus
Siz
e,%
-
12 WOJTANOWICZ, SMITH, NOVAKOVIC, CHIMMALGI, NEWMAN, DUSTERHOFT,
& GAHAN SPE 77752
Figure 5 Shearing rate in plug flow
Figure 6 Cement Pulsation Design Software
12.25X9.625 Annulus; PV = 80 cp: YP = 26 lbs/100 sq ft.
0
100
200
300
400
500
600
0 1 2 3 4 5 6 7 8 9 Velocity, ft/sec
Exa
ApproxiS
hear
Rat
e, 1
/sec
-
SPE 77752 CEMENT PULSATION TREATMENT IN WELLS 13
Figure 7 Algorithm of Cement Pulsation Design Software
Figure 8 Example Well Schematic
-
14 WOJTANOWICZ, SMITH, NOVAKOVIC, CHIMMALGI, NEWMAN, DUSTERHOFT,
& GAHAN SPE 77752
Figure 9 Treated depth vs. time for constant-pressure
pulsation
Figure 10 Top displacement amplitude vs. time for
constant-pressure pulsation
-
SPE 77752 CEMENT PULSATION TREATMENT IN WELLS 15
Figure 11 Treated depth vs. time for controlled-depth
pulsation
Figure 12 Top displacement amplitude vs. time for
controlled-depth pulsation
-
16 WOJTANOWICZ, SMITH, NOVAKOVIC, CHIMMALGI, NEWMAN, DUSTERHOFT,
& GAHAN SPE 77752
Figure 13 Top pressure vs. time for controlled-depth
pulsation