C E M E N T I N G

O P E R A T I O N S

CEMENTING THE CONDUCTOR CASING ANNULUS IN AN OVERPRESSURED WATER FORMATION

Special techniques are required in deepwater offshore areas to cement unconsolidated formations with overpressured-water flows. Overpressured-water sands cause saltwater from the formation to flow into the cement sheath. Prevention of this flow is critical because remedial work to stop flow and repair damage is far more expensive than proper preventive steps. Containment of the overpressured-water formation is complicated by the presence of low-fracture-gradient zones that can fracture and cause loss of mud and cement returns. A successful cementing job depends on three things: (1) hole preparation before running the casing and cementing the annulus, (2) proper slurry weight to control formation-water influx and prevent formation fracture, and (3) short slurrytransition period at cool temperatures.
HOLE PREPARATION

ical 2620-in. casing annulus is less than 30 ft/min and pipe movement cannot be performed easily, the fluid properties of the wellbore fluids are critical. Mud-Based Spotting Fluid. A spotting fluid is the last fluid pumped into the well before operators remove the drillpipe to run casing. In the past, operators used prehydrated bentonite muds with no fluid-loss-control agents. This mud has progressive gel-strength development, creating a fluid that is difficult to displace during the cementing process. Lowering the gel-strength profile of the spotting fluid with a starch fluid-loss-control agent improves cement displacement of the mud. This additive also lowers filter-cake development. Gel-strength values at 10 seconds, 10 minutes, and 30 minutes were 5, 10, and 12 lbf/100 ft2, respectively. The improved spotting fluid should be placed in the hole before running casing because the annular velocity is highest when the casing string is run and drillpipe and casing movement help break up pockets of gelled mud. This helps replace the sweep and spud mud with a drilling

fluid that is easier to displace during the cementing process. Settable Spotting Fluid. A hydraulic material that improves the sealing of the annulus can be added to the fluid system. The hydraulic material will set shortly after the cement slurry is placed or after approximately 10 days. Any filter cake or fluid not removed by the cement will solidify to an initial set condition. If any settable spotting fluid remains after the cement is placed, the spotting fluid will solidify and provide a near-complete seal of the annulus. This becomes important because of limited casing centralization. Low annular velocities and lack of centralization promote cement channeling on the narrow side of the annulus. Foam Sweep. A final sweep of the hole is made with a foamed sweep before the settable spotting fluid is placed. The foam sweep provides superior cuttings removal, improved gelled-mud displacement, and fluid-loss control. Because foaming a fluid increases the viscosity of the base fluid, a properly designed foam sweep will have higher rheology values than any single-

Pipe movement increases the amount of drilling fluid moving in a typical annulus and supplies a mechanical means for breaking the gels of immobilized mud that otherwise would not be broken by fluid flow. Because the annular velocity in a typ-

Fig. 1Cross sections of nitrogen-foamed Portland/ultrafine Portland cement.

834

AUGUST 1997

C E M E N T I N G

O P E R A T I O N S

phase fluid that can be pumped. A relatively thick fluid commonly used for sweeping the hole can be pumped and then foamed downstream of the pumps.
CEMENT SLURRY DESIGN

Slurry Density. The hydrostatic pressure of the slurry must be equal to or slightly greater than the formation pressure, or the formation fluids will flow into the wellbore. Weak formations will fracture when excessive hydrostatic pressure is applied. Slurry density should fall within an approximately 1-lbm/gal window between the pore and fracture pressures. Cement slurries have a transition period during which the gel strength increases because of the hydration process. During this time, the hydrostatic pressure on the formation is reduced. If a large reduction of hydrostatic pressure occurs, the formation fluids could enter the setting slurry. The slurry needs to be designed to maintain sufficient hydrostatic pressure during this transition period. Large-Scale Influx Testing. Three largescale tests were conducted to determine how different cement slurries respond to the influx of water. These tests simulated the downhole conditions of the conductor string. The test apparatus comprised two strings of 2-in. tubing 500 ft in length. One string had pressure ports at 125, 250, and 375 ft. The cement was pumped down one string and returned through the string with pressure ports. A 1,500-psi pressure was placed at the top of the tubing to simulate the seawater and slurry-column hydrostatic pressure. Throughout the tests, water influx was simulated by a constant 9.3lbm/gal gradient through a porous entry point at 475 ft. Testing continued until the cement set. Cross sections were cut and examined for evidence of channeling. The water-extended slurry showed the worst channeling of the three slurries tested. The water flowed into the slurry column and migrated upward as far as 75 ft above the entry port. The water influx displaced enough slurry volume that the test had to be stopped because of slurry entering the pressure-measurement devices at the surface. The water channeling of the incompressible Portland/ultrafine Portland slurry was limited to the location directly above and below the influx port. This slurry was designed to have a short transition time (30 minutes) and absorbed the water influx without being diluted like the slurries that were made less dense by water. The foam slurry did not show any channeling or any displacement by water influx
838

(Fig. 1). During the test of the foam slurry, the pump supplying water pressure to the influx port did not pump any water into the column when set at the 9.3-lbm/gal gradient. It was checked periodically during the test to ensure that the port was not blocked by the slurry. The pump pressure was increased three times during the test to allow a small amount of water to enter the port. This test showed that the foam slurry would prevent water influx better than the incompressible Portland/ultrafine Portland slurry by maintaining a relatively constant slurry-column pressure even when the filtrate volume was removed.
RECOMMENDATIONS

Shortening the transition period of the slurry will limit the amount of water influx into the setting slurry. In shallow-water flows, any influx of water will dilute the slurry. The water influx will change the slurrys properties and lengthen the transition time. Large-scale testing has shown that waterextended slurries have setting times of approximately 2 hours, allowing influx and migration of water. A special blend of Portland and ultrafine Portland cement with additives has a 30-minute transition period, which allows only a minimal amount of water influx. Foamed Portland/ultrafine Portland cement exhibited superior water-control characteristics. It has short transition and set times at 55F . The density of a foamed cement can be adjusted by varying the gas content, which is important when formation pressures are not known before the well is drilled. Test data indicate that the pressure of a foamed slurry remains almost constant even while it progresses through its transition period. This is caused by expansion of the gas used to foam the cement as the cement volume decreases. The two-phase slurry helps limit the volume reduction by controlling the slurrys filtrate loss. This cannot be achieved with a noncompressible slurry.
EXAMPLE PROCEDURE

2. Give the data to the cementing company to allow engineers to design the drilling fluid and cement job to control flow from the sands. 3. Drill the pilot hole to approximately 2,000 ft. 4. Use drilling-fluid sweeps with lowfluid-loss properties when drilling through and below the sands. 5. Drill out the pilot hole to 26 in. 6. When total depth (TD) is reached, displace the hole contents with a low-fluidloss drilling fluid and short trip the bit to drive the pipe shoe . 7. Run in the hole to TD and displace the hole with a foamed sweep and a low-gelstrength drilling fluid or a settable spotting fluid. Keep the hole static with a fluid approximately 0.5 lbm/gal greater than sand pore pressure. 8. Remove the drillpipe and run casing. 9. Pump one or two inner-string volumes of drilling fluid to detect operational problems. 10. If the drilling fluid and cement are not compatible, use a spacer with a density greater than the drilling fluid density. 11. Cement the casing through the inner string.
CONCLUSIONS

Operators have successfully used slurries consisting of nitrogen-foamed cement to help reduce hydrostatic pressure on weak formations. Accelerators, such as ultrafine cement, are added to these slurries to hasten the onset of hydration in the 40 to 60F water temperature. Drilling fluid with low fluid loss and gel strength should be used to stabilize the 26-in. hole. A starch-based fluid or settable spotting fluid is recommended because they do not require centralized casing. Hole preparation is just as important as slurry selection in controlling overpressured-water zones. Foamed drilling sweeps can develop the shear stress necessary to erode the partially dehydrated drilling fluid but not the filter cake. This article is a synopsis of paper OTC 8304, Cementing the Conductor Casing Annulus in an Overpressured Water Formation, by James Griffith, SPE, and Ronnie Faul, SPE, Halliburton Energy Services Inc., originally presented at the 1997 Offshore Technology Conference held in Houston, 58 May. Please read the full-length paper for additional detail, illustrations, and references. The paper from which the synopsis has been taken has not been peer reviewed. Copyright 1997 Offshore Technology Conference.
AUGUST 1997

The following procedure was followed to set 20-in. casing through a water-flow zone. A riser was not available and there were no geophysical data to help operators avoid major-water-flow sands. Drive pipe is set to approximately 200 ft below the mud line (BML), and 26-in. casing is set approximately 1,000 ft BML. 1. Drill a 9- to 12-in hole and obtain data to locate the depth, thickness, and pore pressure of the sands. Use seawater and gelled drilling-fluid sweeps.

C E M E N T I N G

O P E R A T I O N S

CEMENTING IN AREAS PRONE TO SHALLOW SALTWATER FLOWS

BACKGROUND

Shallow-saltwater flows exist in the Gulf of Mexico on the transitional slope between the continental shelf and deeper water in water depths from 1,000 to 7,000 ft. Isolated flows have been reported in water as shallow as 500 ft. The sands containing the water flows are called turbidites because they were deposited by turbidity currents. Turbidity currents are tongue-like masses of debris-laden water that flow downslope beneath clearer water because of their greater density. These currents are generated by slumping along slopes on which rapid sedimentation occurs. This slumping may occur when semifluid deposits become unstable on inclined surfaces. Turbidites are typically deposited as graded-sediment beds with the coarsest particles on the bottom. Successive turbidity currents passing over previous deposits may lay down additional graded beds without disturbing the relatively fine upper surface of the previous layers. This results in undercompaction and overpressuring of the sediments. Most of the overpressured sands have been encountered between 500 and 2,000 ft below the mudline (BLM). The pore pressure in these sands ranges from 9.0 to 9.6 lbm/gal equivalent mud weight (EMW). Pore pressures are determined by use of pressure-while-drilling sensors as well as from the density of the kill mud required to control the flow.
DRILLING PRACTICES

tural integrity of the wellbore is critical. Proper casing-point selection, casing-seat isolation, and casing structural support are critical to the life of the well. The combination of very soft unconsolidated sediments and shallow-water flows make the possibility of casing collapse, buckling, and wear a major concern.
CEMENTING

The shallow hole sections across the flow sands are usually drilled without a riser. Because returns are taken at the seafloor, the use of weighted drilling fluids is impractical. These sections are drilled with seawater. Viscous-gel sweeps are pumped at regular intervals to remove cuttings from the hole. Once the casing point is reached, the hole is loaded with kill mud before running casing and cementing. The mud must have good fluid-loss control to lay down a thin impermeable filter cake. It should be viscous enough to remove residual cuttings yet thin enough to allow effective removal during cementing. When wells are drilled in deepwater shallow-flow areas, the struc840

The objectives for cementing in shallowflow areas are threefold. The first is to achieve a competent hydraulic seal that will not allow migration of fluids, especially saltwater flows behind the pipe. Second is to give structural support to the pipe so it will resist buckling and casing wear. Third, to provide long-term durability for the life of the well. Successful cementing requires dealing with problems of weak unconsolidated formations, overpressured water sands, narrow margins between fracture and pore pressures, large-diameter washedout openhole sections, and low bottomhole temperatures. Both the conductor and surface casings are cemented by use of an inner string of drillpipe. This is done to minimize displacement volumes, cement contamination, and amount of seawater that must be circulated from the inside of the pipe once the casing is landed. The casing is hung off at the mudline, and returns are taken through cementing ports at the seafloor. In some instances, the surface pipe may be cemented with a riser in place and returns are to the rig floor. The well must be static before and during cementing. The kill mud in the hole must be the correct density to control saltwater flows without breaking down the weak unconsolidated sediments. Fluids pumped during the cement job must also be of the proper density to maintain well control. Mud Removal. The same mud-removal principles that apply in normal cementing situations also apply in deep water. Good casing centralization is difficult, but achievable. Casing sizes and subsea-wellheadequipment sizes may have to be changed from what is normal. Openhole size can be

optimized to enhance centralization and mud removal. Studies have shown that a minimum increase of 10% in density for each successive fluid is ideal for effective displacement of one fluid by another in an annulus. The narrow difference between pore and fracture pressures makes this difficult to achieve. The fluid-density differential should be maximized with a kill mud with the lowest density possible and a lead slurry density as high as possible. Cement-System Performance. Cementing in deepwater shallow-flow areas requires special attention to cement-system design, particularly for the lead cement. The cement slurry must exhibit certain hydraulic properties for proper placement in the wellbore while the set cement sheath must possess certain mechanical properties for long-term zonal isolation and structural support. In the fluid state, consideration must be given to stability, fluid-loss, rheology, thickening time, and critical hydration period (CHP). Once the cement is set, acceptable compressive strength, permeability, shear-bond strength, and ductility are required. Obtaining the desired performance properties in the deepwater environment is a challenge. The temperature at the seafloor is from 35 to 45F and the bottomhole static temperature may be as low as 50F . The density of the lead-cement system is from 11.0 to 12.0 lbm/gal because of low fracture gradients. Many of the required hydraulic and mechanical properties are difficult to achieve under these conditions. Slurry Stability. The required slurry stability is defined by no free water and no significant settling or sedimentation. If the slurry has excessive free water, channels will form that may result in a loss of zonal isolation. As water is removed from the slurry the pressure in the cement column will drop, possibly permitting an influx of reservoir fluids. Settling will cause density differentials throughout the cement column that may result in insufficient hydrostatic pressure to maintain well control.
AUGUST 1997

C E M E N T I N G

O P E R A T I O N S

Gel strength (lbf/100 ft2)

properties because of the flat displacement profile at the leading edge of the foamed slurry. Foamed cement has the logistical advantage of variable density in a single blend. This is important when unplanned changes in mud density occur just before cementing. The density of foamed cement can be varied over a range of 5 to 6 lbm/gal with minimal effects on the properties of the cement.
C A SE HISTORY

Fig. 1Gel-strength development vs. time for cement slurries.

Fluid Loss. Fluid-loss control, both dynamic and static, is an essential property. When excessive amounts of fluid are lost from the slurry during placement (dynamic fluid loss), the slurry density increases, which may result in lost circulation and may affect properties such as thickening time. Loss of fluid from the slurry after placement (static fluid loss) will lead to a slurry volume reduction and drop in interstitial pressure that can allow reservoir fluids to enter the slurry. CHP. The CHP is a period of time that begins when the slurry no longer transmits hydrostatic pressure that overbalances the pore pressure of the formation and ends when the slurry has developed enough cohesive strength to prevent entry and flow of reservoir fluids into the annulus. Fig. 1 illustrates the evolution of cement-slurry gel strength during the CHP . The lower critical limit, tc, is the time when the gel strength is equal to the critical wall-shear stress (CWSS). CWSS is the gel strength at which the cohesive forces between the cement slurry, the wellbore walls, and the casing become strong enough to cause the hydrostatic pressure to decline to a pressure equal to the formation pore pressure. CWSS can be calculated on the basis of the geometry of the cemented portion of the wellbore, the pressure at the top of the cement column, and the height and density of the cement column. tf, the final time, is the time when the slurry is effectively impermeable to fluid entry and migration through the cement matrix. tf must be determined experimentally. Experimental studies with
AUGUST 1997

a cement hydration analyzer indicate that gel strength values as high as 10,000 lbf/100 ft2 may be required to prevent fluid influx and flow through a cement slurry. Fluid migration can be controlled by shortening the CHP . CHP can be minimized by increasing the CWSS, decreasing the gel-strength value required to stop fluid entry and migration, or by increasing the slope of the gel strength vs. time curve. In Fig. 1, increasing the CWSS by 10 units would have approximately the same effect in shortening the CHP as decreasing tf by 1,000 units. Changing the CWSS is not practical in deepwater shallow-flow situations. tf can be decreased by optimizing the slurry formulation with anti-fluidmigration technology. The CHP can be shortened by increasing the slope of the gel-strength curve. This is accomplished by chemically altering the composition of the slurry with special gel-strength modifying additives. Compressible Cements. Compressible cements can be used to control saltwater flows. The compressed gas in the cement maintains a matrix pressure in the slurry greater than the pore pressure of the flowing zone as the cement undergoes the CHP . The large volume of gas in the slurry will also help compensate for slurry-volume reductions caused by fluid loss and volumetric shrinkage. Compressible cements used for gas-migration control are foamed cements or in-situ gas-generated cements. Foamed cements exhibit good fluid-loss control and compressive-strength development at low densities and temperatures. Foamed cement has good mud-removal

Two offset wells had been drilled in the Gulf of Mexico block in 3,665 ft of water. Both offsets experienced serious water flows during drilling, and one had to be abandoned because of uncontrolled flow. A 26-in. conductor casing was set at 985 ft BML approximately 165 ft above the highest sand indicated by seismic data and was cemented with foamed cement. A pilot hole of 9 in. was then drilled with seawater and gel sweeps to a depth of 1,853 ft BML. Four separate flowing sands were drilled from 1,150 to 1,565 ft BML with a total net thickness of approximately 100 ft. Static-flow checks indicated substantial flow from all four sands. Pore pressure for the sands was approximately 9.2 lbm/gal EMW. The hole was loaded with 12.0lbm/gal kill mud and was static. The EMW at the top sand at 1,150 ft BML was 9.2 lbm/gal with the 12.0-lbm/gal mud in the hole. The hole was opened to 26 in. and reloaded with kill mud, and 20-in. casing was run. The 20-in. surface casing was cemented by use of 60 bbl of 12.3 lbm/gal spacer followed by 1,100 sacks of foamed right angle set (RAS) lead at 12.5 to 13.4 lbm/gal, 1,585 sacks of foamed RAS tail at 15.0 lbm/gal, and 340 sacks of nonfoamed RAS tail at 15.8 lbm/gal. The cement slurry was mixed with a process-controlled mixer, and the nitrogen and surfactants were pumped at the proper ratio with a computer process-control system. There was no flow after cementing. This article is a synopsis of paper OTC 8305, Successful Cementing in Areas Prone to Shallow Saltwater Flows in Deepwater Gulf of Mexico, by D.A. Stiles, SPE, Dowell, originally presented at the 1997 Offshore Technology Conference held in Houston, 58 May. Please read the full-length paper for additional detail, illustrations, and references. The paper from which the synopsis has been taken has not been peer reviewed. Copyright 1997 Offshore Technology Conference.
841

C E M E N T I N G

O P E R A T I O N S

IMPROVED PRIMARY CEMENT JOBS WITH A UNIQUE SPACER DESIGN

A good primary cement job across a productive horizon is considered one of the most critical points at the end of the drilling stage and the beginning of the completion stage of an oil well. In the U.S. Gulf of Mexico (GOM) offshore and gulf coast (GC) onshore, poor cement jobs have been very costly, especially in areas where formations have a distinct oil/gas/water contact and where complete isolation is desired across a pay interval. Lack of zonal isolation has resulted in expensive squeeze repair operations, reduced production rates, premature watering out, failure of stimulation treatments, and other problems that ultimately increase operating costs and cause loss of reserves. Cement bond logs (CBLs) are used to determine the quality of the cement. Contamination lowers the compressive strength of the cement. The decrease in strength causes the amplitude of the CBL to be higher than expected. This increase in amplitude can be confused with a cement channel. A cement evaluation log is used to examine 5 to 60 segments around the casing to determine whether the CBL results are caused by contamination or channeling. One method to identify contamination uses a statistical variation of the impedance or amplitude curves to determine that cement exists behind the casing. If the cement around the casing is contaminated but has uniform quality and a compressive strength greater than 200 psi, it has adequate strength for zonal isolation. When remedial cementing is required, some jobs have placed additional cement behind the casing but it often does not fully cover the casing circumferentially or extend more than a few feet. In other cases, cement could not be squeezed behind the casing.
GOOD CEMENT DESIGN

Achieving a good primary cement design involves the cooperation of many people and selection of the correct cement for the specific well conditions. Many of the problems in the GOM and GC were results of gas influx. The primary cement should contain the necessary additives specified in the literature, and testing should be conducted with attention to the cement char842

acteristics that prevent gas influx. As cement sets, it undergoes a loss of hydrostatic pressure, which can allow gas to migrate into the wellbore and cause contamination. Important cement characteristics include low fluid loss, an adequate yield point to displace mud without fingering, fast right-hand-set transition times, and rheology control. Traditionally, cement spacer fluid and spacer-displacement techniques have not received sufficient attention because their importance was not understood. Inefficient displacement techniques have resulted in bypassed mud, viscous fingering, poor wellbore cleaning, and excessive mud/ cement mixing. Mud and cement mixing can cause hydrostatic pressure loss in the column and allow annular gas flow into the wellbore. Spacer technology, combined with cement design, pipe centralization, and proper job execution, has resulted in good downhole placement of cement slurry with minimal mud contamination and excellent cement bonds. Uncentralized casing strings can cause channeling and affect mud displacement and cement placement. Guidelines exist for centralizer spacing based on wellbore deviation. One practice places two solidbody centralizers on each joint of casing for 200 ft above and below a pay interval. Often one centralizer per joint is run across depleted sands to prevent differential pipe sticking. Experience has shown pipe movement (reciprocation), coupled with good centralization, is beneficial for obtaining good mud displacement when conditioning the hole before cementing. Use of a mandrel or mudline suspension hanger will hamper reciprocation, and reciprocation is impossible if a liner hanger is set before cementing. Mud conditioning to lower yield point and plastic viscosity is essential to achieving a good cement job.
DISCUSSION

pumped to displace and clean out the drilling mud in the wellbore. The spacerfluid system typically includes a weighted water-based spacer containing surfactants, a viscous polymer gel, and a final wash fluid that can be a simple aqueous surfactant wash for water-based mud (WBM) or an organic solvent coupled with a buffer brine for oil-based mud (OBM) and synthetic-based mud (SBM) The keys to displacement-cleanout success are the ability to minimize the interface volume and provide superior cleaning by (1) controlling the density and viscous transition between adjacent fluids, (2) maintaining excellent solid suspension in the interface, (3) using highly efficient specialty chemicals to promote both compatibility between fluids and a fast and effective cleaning of the drilling mud. In OBM displacement, most of the OBM is displaced first in a piston-like manner during the weighted-spacer and viscousgel stages. This is followed by turbulent cleaning of residual oily contaminants in the solvent-wash stage. The weighted water-based spacer was optimized by including an efficient surfactant system. The improved design of this weighted spacer, Arco SL Spacer, has been field tested and used successfully in the displacement cleaning of OBM, WBM, and SBM from the borehole. The efficient displacement cleanout of drilling mud eliminated a common contamination problem in downhole transport and placement of the cement slurry, providing a better cement bond during cementing operations. Spacer Design. The Arco SL Spacer typically comprises a base aqueous fluid, a blend of viscosifiers, a copolymer dispersant, and sugar-lipid nonionic surfactants. The spacer is a pseudoplastic fluid with a relatively flat shear-stress profile over a wide range of shear rates. Its effectiveness is relatively unaffected by the existence of eccentric annuli in the borehole. A nearconstant shear-stress profile results in a more uniform distribution of hydraulic force on the wellbore surfaces, enabling the removal of partially dehydrated or gelled drilling mud and solids from the wellbore.
AUGUST 1997

Arco has developed effective wellborecleanout technology that combines an effective displacement technique with differential spacers. A sequence of small volumes of specially designed spacer fluids is

C E M E N T I N G

O P E R A T I O N S

Temperature has little effect on the viscosity of the spacer. This minimizes the thermal-thinning effect on the fluid, which can adversely affect its displacement efficiency as well as its solid suspension properties because of thinning of low-shear viscosities. The viscosifier blend, which includes a bentonite gel, provides a compensating effect on fluid viscosity as temperature increases during the downhole transport of the spacer fluid. In spacer designed to clean WBM, the high-performance copolymer dispersant provides compatibility between drilling mud and cement slurry should they become intermixed, preventing severe gelation or flocculation, which can completely plug the annular spaces. The dispersant also is a stabilizer for bentonite at high temperatures. For spacer designed to clean OBM and SBM, miscibility between the mud and water-based spacer is achieved by use of a nonionic surfactant blend of sugar-lipid surfactants with or without linear alkyl ethoxylated alcohol. The surfactants are designed to form a microemulsion with ultralow interfacial tension over a wide

range of temperatures when mixed with the OBM or SBM. This aids in cleaning the oily mud residue and filter-cake deposits. The formation of a microemulsion prevents the creation of a highly viscous emulsion that tends to be left behind and becomes the source of cement contamination. The use of nontoxic, biodegradable, and highly hydrophilic sugar-lipid surfactants not only improves the water-wetting properties of surfaces, and therefore the cement bonding, but also minimizes the impact on the environment. The hydrophilic surfactants promote the hydration of viscosifiers, which aids in the development of static gel strength for solids suspension. Both the dispersant and surfactants in the spacer provide additional benefits in the removal of partially dehydrated or gelled drilling mud because of their dispersing properties.
CONCLUSIONS

has resulted in good downhole placement of cement slurry with minimal mud contamination and excellent wellbore displacement of successive fluids. It has not only reduced the need for remedial cement work and reduced rig time for displacements but has also simplified cement-bond evaluation. This spacer technology has been successfully applied in WBM, OBM, and SBM systems. This article is a synopsis of paper SPE 36486, Improved Primary Cement Jobs Through the Use of Unique Spacer Design Technology: Gulf of Mexico Case History Study, by J.P . Schumacher, SPE, R.W. Bell, and S.E. Morrison, Vastar Resources, Inc., and A.F. Chan, SPE, and R. Wydrinski, SPE, Arco E&P Technology, originally presented at the 1996 SPE Annual Technical Conference and Exhibition held in Denver, Colorado, 69 October. Please read the full-length paper for additional detail, illustrations, and references. The paper from which the synopsis has been taken has not been peer reviewed.

The Arco SL Spacer has been tested extensively in a number of offshore wells in the GOM and produced outstanding results. Multiple field examples show that this cleanout technology, combined with pipe centralization and proper job execution,

AUGUST 1997

843

C E M E N T I N G

O P E R A T I O N S

NEW TECHNOLOGY IMPROVES CEMENT-SLURRY DESIGN

A promising geothermal concession is located in a tea plantation on the island of Java. A drilling project was undertaken to evaluate and harness this resource for geothermal electricity generation. The program used two slimhole rigs to drill appraisal wells to establish the potential of the field. Companies are being faced with environmental concerns throughout the world. The requirement to reduce the size of drilling sites in environmentally sensitive areas has necessitated a change in the way drilling operations are conducted. Cementing companies historically have needed a large work area for bulking and mixing equipment. Cement dust and wash-down water emissions presented other environmental concerns. Geothermal wells present the most severe conditions to which cements are exposed. As a result, their performance requirements are among the most stringent. Geothermal cements are usually designed to provide at least 1,000 psi compressive strength and no more than 1.0-md water permeability. While casings with tight annular clearances require that good cementing practices be observed, they also create conditions that demand much greater care and control in slurry and procedure design than regular casing cementation. Free-water and thickening-time requirements are similar for geothermal and slimhole conditions, but the use of perlite and silica flour complicate the rheology required for geothermal wells.
LIQUID -CEMENT PREMIX

cements. The LCPs are activated and modified if necessary to yield a finished slurry with suitable properties for well cementing. An activator is added immediately before pumping the slurry into the well. This activator reinitiates the chemical hydration and allows the slurry to develop compressive strength after a predetermined pumping time. Properties of LCP . Before activation, the LCP behaves as a weighted mud. It is stored as a slurry with a density of 16.0 lbm/gal in the form of a stable suspension. Agitation is not usually required for LCPs, but stirring is beneficial for slurries with a high-silicaflour loading to ensure that there is no settling and to offset any slight thixotropic tendency. After activation, the LCP behaves as a conventional cement slurry with an excellent right-angled set. As a set cement, it develops normal levels of compressive strength and other mechanical properties. Slurries with short thickening times can develop early compressive strengths of 50 to 100 psi in 3 hours. A density range of 11 to 21 lbm/gal can be obtained by adding extenders, or weighting agents. Chemistry of LCP . When conventional Portland cement is added to water, a chain of complex chemical reactions is set in motion. In a matter of hours, the mixture is transformed irreversibly into a hard product with considerable mechanical strength and low permeability. The mechanical strength is a function of the increase in surface area of the cement caused by the formation of voluminous amorphous hydrates coupled with crystalline deposits of portlandite (calcium hydroxide) and ettringite (calcium sulfoaluminate). The setting of Portland cement is a solution/precipitation process in which somewhat soluble minerals dissolve to produce low-solubility hydrates, which then precipitate from supersaturated solutions. The nucleation and growth of these hydrates is an essential feature of the setting process. In the case of LCP , the combination of set-retarding and conditioning agents slows the forward-hydration reactions to an imperceptible level. The solution com-

position of the interstitial fluid is changed significantly with respect to certain key ions, and the nucleation and growth of hydrates is inhibited. Because of the changes, interparticle forces between cement grains are reduced and sedimentation of the solids must be prevented. This is accomplished by the conditioning agents in the LCP formulation.
APPLIC ATION

A central cement-mixing and -storage facility was set up outside the environmentally sensitive area. The LCP would be mixed and tested to ensure that it met the design specifications. It would be transported to the wellsite in a purpose-built liquid-cement unit. The preliminary well program called for three sizes of casing to be cemented back to surface with lightweight lead slurries used on the two deeper strings.
L ABORATORY TESTING

Liquid-cement premix (LCP) is a storable cement slurry that can be kept in a liquid state indefinitely and made to set as and when required. It consists of a premixed slurry of Portland cement (or other cementatious material) containing set-retarding and conditioning agents, with water as the carrier fluid. This storable slurry can be tailored to have a storage life of several days to more than 6 months, and it can be formulated with standard oilwell cements. American Petroleum Inst. Classes G, H, C, and A cements have all yielded satisfactory LCPs, and it is also possible to prepare storable slurries based on blended
844

An extensive laboratory testing program was undertaken to design the optimum LCP for all the programmed casing strings. The original 16.0-lbm/gal LCP formulation met, and in some cases exceeded, the design criteria under laboratory conditions. The addition of a liquid additive package (LAP) also produced acceptable lightweight slurries, enabling the cementing of all the programmed casing strings and any lost-circulation zones encountered during drilling with a single LCP formulation. Where jobs required lead and tail slurries, sufficient LCP for both stages was transported to location in one tank. LAP to mix the required quantity of lead slurry was in another tank. Before the job, the required quantity of LCP was added to the LAP to produce a nonactivated, lightweight LPC that was stored for up to 5 days. The storage life of the LCP was designed to exceed 21 days (Fig. 1). Because of the storage time and the requirement that thickening times be relatively short, two activators were used to reawaken the liquid cement.
FIELD PILOT TESTING

To confirm the laboratory test results, a batch of LCP was mixed in the field. Field
AUGUST 1997

C E M E N T I N G

O P E R A T I O N S

FANN 35 Reading R1 B1 F1

DAYS

RHEOLOGICAL READINGS - FANN 35 (R1, B1, S1)

6 7 8 9 10 11 12 13 14 15

rev/min

rev/min

16

17

18

19

20

21

22

23

Fig. 1LCP storage life.

testing identified two phenomena that were not experienced during laboratory testing. 1. The viscosity of the field-mixed LCP was slightly lower than the rheology readings for the laboratory samples. 2. The field-mixed LCP exhibited a settling tendency when the system was exposed to severe vibrations. This test was performed in an effort to duplicate the anticipated vibrations that the LCP would experience during transportation. To reproduce properties exhibited in the laboratory, additive concentrations were adjusted and the slurry was mixed at a faster rate to reduce recirculating time. The settling tendency was attributed to the high-silica-flour loading because severe vibration had no detrimental effect on LCP without silica flour. The LCP was agitated during transportation to overcome the effects of vibrations.
FIELD OPERATIONS

lished, 20 bbl of LCP was circulated in the batch mixer and the activator added. After 5 minutes, the slurry was pumped and displaced with full returns to surface. Three hours after the cement was put in place, the samples of the activated LCP were hard enough to support the weight of the casing. The landing string was backed out, and wellhead installation work begun. When cementing operations required a lightweight lead slurry and a heavier tail-in system, additional storage tanks were used. Fluid-loss additive and perlite were mixed in water to form an LAP . LCP was added to the LAP while the casing was run to raise the density to meet design specifications. When the casing was in place and circulation established, activator was introduced and the cement slurry pumped and displaced. More than 30 cement jobs have been completed successfully with LCP in Indonesia.
CONCLUSIONS

2. The use of LCP can eliminate problems associated with well cementing in environmentally sensitive areas. 3. LCP allows simplification of cementmixing equipment and improves wellsite quality control and logistics. 4. Mixing time is eliminated, so high-rate cement-slurry placement is practical when LCP technology is used. 5. Large LCP jobs are now a realistic alternative to conventional cementing techniques. This article is a synopsis of paper SPE 36973, New Technology Improves Cement Slurry Design, by J.W. Anderson, SPE, A.I. Buchanan, SPE, and A. Susanto, P .T. BJ Services Indonesia, originally presented at the 1996 SPE Asia Pacific Oil and Gas Conference held in Adelaide, Australia, 2831 October. Please read the fulllength paper for additional detail, illustrations, and references. The paper from which the synopsis has been taken has not been peer reviewed.
845

One day before the job was planned, 35 bbl of LCP was transported to location. When the casing was landed and circulation estab AUGUST 1997

1. An LCP formulation will meet, and in most cases exceed, the design criteria for an acceptable geothermal cement slurry.

C E M E N T I N G

O P E R A T I O N S

RECIPROCATING CEMENT SLURRIES AFTER PLACEMENT

The construction industry routinely vibrates concrete slurries to improve the quality of the set concrete. Several attempts have been made to adapt this technology to well cementing. Casing has been vibrated in contact with well cement slurries in the laboratory and in full-scale tests. Largescale hydraulic equipment has been constructed to support and vibrate the casing string directly. The method described in this paper uses very simple and inexpensive equipment to introduce pulses of compressed air or water directly into the annulus above the slurry. The annulus serves as a wave guide to transmit pressure pulses efficiently through the slurry deep in the well. The objective of reciprocating, or vibrating, the slurry is to improve the bond between the casing and cement by preventing the influx of gas into the cement slurry before it sets. After a cement slurry is pumped, it undergoes a hydration reaction that transforms it from a liquid slurry to a solid. During this transition it becomes a gel and its volume decreases slightly. The combination of gelation and shrinkage progressively transfers the weight of the slurry onto the surfaces of the casing and the wellbore, resulting in a decline in the hydrostatic pressure (Fig. 1). Gas influx can occur when the hydrostatic pressure is less than the formation pore pressure. This pressure decline has been measured by attaching pressure transducers to the outside of the casing. Hydrostatic pressure has been increased at shallow depth by water injection. At greater depths, the time required for the hydrostatic pressure to decline to the pore pressure was about twice that of wells without applied pressure. Relatively low pressure applied to the annulus at the surface can restore the hydrostatic pressure at substantial depths. Operators have tried injecting water continuously into the annulus at a low rate to maintain a constant pressure at the surface for several hours after cementing. There is no evidence that this practice improved cementing operations. It is necessary to apply and release the pressure periodically to maintain the hydrostatic pressure and to move the cement particles continually to prevent them from forming the short-range interactions that cause gelation.
FIELD TESTING

An idle well in the Humble field near Houston was used to test the equipment. It was 2,880 ft deep with 6-in.-inside-diameter casing and completely filled with water. Pressure pulses from water-pulse generators could be easily transmitted down the well and reflected back. When the frequency was adjusted to coincide with the travel time, resonance occurred. Under resonant conditions, peak pressures could exceed the pressure of the pulses from the waterpulse generator. A 5-ft-high, 6-in.-diameter plastic tube was attached to the top of the well to permit observation of the motion of the water surface. An air-pulse generator produced a 20-in. water displacement. When its frequency was adjusted to resonance, the displacement was greater than the 5-ft height of the plastic tube. Transmission of Pressure Pulses. To demonstrate that pressure pulses could be transmitted efficiently through cement slurries, a bridge plug was set at a depth of 300 ft inside 4.5-in. casing for the surface plug of a well abandonment in the Mabee field near Midland, Texas. Neat American Petroleum Inst. Class H cement slurry was circulated to the surface. Single-shot pressure pulses were applied to the cement slurry at the surface and were monitored at depths of 0, 100, 200, and 300 ft by lowering a specially constructed wireline log through the slurry. Fig. 2 shows the resulting pressure at each depth. An electronically controlled water-pulse generator was used for this test. It was used in a singleshot mode powered by nitrogen gas at a pressure of 300 psi. Vibrating Slurries in Wells. Tests were performed on shallow vertical oil wells with no gas-migration problems. These wells were drilled in the North Concho (Queen) field near Odessa, Texas. The 85/8-in. surface casing was set at 1,500 ft and drilled out to a true vertical depth of approximately 4,700 ft. The 51/2-in. production casing was cemented to surface. The lead slurry was 12.8-lbm/gal 35/65 Poz/Class H cement with 6% bentonite and the tail slurry was 14.2-lbm/gal 50/50 Poz/Class H ceAUGUST 1997

8,754 ft

5,488 ft 3,636 ft

Time, hours Fig. 1Hydrostatic pressure decline after cementing.

846

C E M E N T I N G

O P E R A T I O N S

Downhole Transducer

100-ft Surface Transducer

Downhole Transducer

200-ft Surface Transducer Downhole Transducer

rating was 120 psi. When tested on the Humble well, this pulse generator would only cycle in resonance with the return time of pulses from the bottom of the well. An improved water-pulse generator was machined from aluminum-alloy halves bolted together to provide an internal chamber. Compressed air or nitrogen was introduced into one end of the chamber to accelerate a pulse of water out of the other end. The water was separated from the gas by a diaphragm manufactured for the first pump. Electronically controlled valves were used to inject and exhaust the gas, and the backpresssure of the water returned the diaphragm to its initial position. It provided a water pulse with a displacement of approximately 0.5 gal and a half-peak width of approximately 0.2 second and had a pressure rating of 400 psi. Air-Pulse Generators. Various air-pulsegenerator configurations were used. They injected and exhausted compressed air directly into the annulus. They had no displacement limitation and provided an average vertical motion of 3.5 ft at 100 psi in the Queen wells. They comprised fast-acting (0.05-second), large-volume pilot-operated air valves with electronic or pneumatic controls. They were either time- or pressureactivated. Time-activated air-pulse generators were used at the rate of 1 cycle/10 sec for these tests. Compressed air from 100 to

120 psi was provided through a 50-ft length of 3/4- or 2-in. hose from trailer-mounted rental air compressors with deliveries of 185 or 375 ft3/min at atmospheric pressure.
DISCUSSION

300-ft Surface Transducer Downhole Transducer

Time, seconds

Fig. 2Transmission of pressure pulses in cement slurry.

During the test in the Mabee well abandonment plug, the amplitude of the pressure pulses and the areas under the curves actually increased with depth. This may have resulted from addition of some energy to the pulse from the relaxation of the slurry as gelation was broken by the pressure pulse. The pressure pulse was not reflected from the bottom of the 300-ft slurry. Reflections might have been absorbed by the same mechanism that amplified them, allowing the pressure pulses to go only one way. The double-peak shape of the pulse did not change with depth. Pressure pulses imparted motion to the cement particles in the slurry in two ways. Narrow pressure pulses from water-pulse generators traveled at the speed of sound, but the relatively small displacement resulted in low amplitudes in all wells except the water-filled Humble well. Air-pulse generators had no displacement limitation, were simpler, and had fewer parts. They could also provide composite pulses with a narrow spike followed by a slower wave. Compressibility Results. Fig. 3 shows the general shape expected for a cement com-

ment with 2% bentonite. The top of the tail slurry was approximately 3,000 ft. The objective of these tests was to maintain the hydrostatic pressure by preventing gelstrength development. This was evaluated by monitoring the compressibility of the cement slurry in the annulus. Both air- and water-pulse generators were used. Cement bond logs (CBLs) were run the full length of the Queen wells between 48 and 72 hours after cementing.
PULSE GENERATORS

The tests were performed with different pulse-generator configurations to determine which worked best in the field. Waterpulse generators provide a small pulse of water in a short time. Air-pulse generators provide a large pulse of air over a longer time period. Water-Pulse Generators. The first tests used a water-pulse generator made from a highly modified 2-in. air-powered dualdiaphragm pump. It had a displacement of approximately 0.5 gal, resulting in a vertical motion of approximately 4 in. in the annulus of the Queen wells. The half-peak width was 0.2 to 0.5 second, and it cycled approximately every 1 to 3 seconds. Its pressure
AUGUST 1997

Fig. 3Behavior of compressibility vs. time.

847

pressibility vs. time curve. The compressibility remained constant until the compressive strength of the cement at the bottom of the well was equal to the pressure of the pulses from the surface. Compressibility declined as the setting process moved toward the surface. Three of the seven Queen wells that were vibrated maintained a relatively constant compressibility for approximately 4 hours. The remaining four lost compressibility rapidly, sometimes within 1.4 hours. Cement Logs. The drilling brine was easily displaced by the cement slurry in the Queen wells. This eliminated the effects of undisplaced drilling fluid on CBLs. CBLs were compared at a depth of approximately 3,700 to 4,000 ft to determine the effect of vibration on the bond between the cement and casing. The CBLs of cement vibrated with an air-pulse generator were better than those of cements not vibrated. Improvement was marginal for cements where the water-pulse generator was used. The CBLs were run without pressurizing the casing and the resulting logs indicated such good bonding that they were outside the normal parameters usually used to display the log printouts. The CBL improvement indicated that vibration prevented the formation of a microannulus as the cement set. The motion of the slurry in the annulus changed from a downward direction to a transverse direction at the zone where the cement was setting. This transverse motion pushed individual cement particles in the slurry toward the casing during each cycle, improving bonding.
CONCLUSIONS

1. A simple, inexpensive technology has been developed that vibrates cement slurries after placement. 2. It keeps the slurry in a fluid state to maintain the hydrostatic pressure and prevent gas migration. 3. It improves the bond between the cement and casing. 4. It eliminates the microannulus to reduce and prevent microannular gas. This article is a synopsis of paper SPE 37619, Reciprocating Cement Slurries After Placement by Applying Pressure Pulses in the Annulus, by J.P. Haberman, SPE, Texaco, and S.L. Wolhart, SPE, Gas Research Inst., originally presented at the 1997 SPE/IADC Drilling Conference held in Amsterdam, 46 March. Please read the full-length paper for additional detail, illustrations, and references. The paper from which the synopsis has been taken has not been peer reviewed.
848 AUGUST 1997

C E M E N T I N G

O P E R A T I O N S

CEMENTING MULTILATERAL WELLS WITH LATEX CEMENT

A multilateral well is a well with one or more branches or lateral sections extending from its main wellbore. The laterals can be openhole or cased hole. When laterals are cased hole, the cement integrity for casing support and zonal isolation is very important. When cementing the lateral sections of multilateral wells, it is important to use a cement with high strength and durability to support the liner throughout the life of the well and to support the lateral section. The cement column is subjected to various stresses when the cemented inner stub is cut. High tensile strength, flexural strength, and crack resistance are required. These properties are necessary to make a clean cut through the cement sheath that does not induce cracks in the cement column.
INTRODUCTION

parallel to the main casing string. During this operation, the cement column is subjected to a high degree of bending and to compressive and tensile stress. Cutting the top of the cement sheath can initiate cracks that can propagate through the cement column. After the liner stubs are cut and retrieved, the cement is the liners only support. Because of this, a cement with highflexural strength and durability is essential. Conventional gas-migration-control cement does not have these properties and will crumple and develop cracks in the cement column, damaging its integrity.
TESTING L ATEX CEMENT

in many respects. Latex-cement slurries showed a marked improvement in flexural strength and fluid loss.
C A SE HISTORY

Latex cement is commonly used for its gasmigration-control property. Its main component is a polymer dispersion of an aqueous solution of styrene and butadiene microparticles stabilized in blends of surfactant commonly known as synthetic rubber. A styrene/butadiene ratio of 60/40 is known to give a good balance of properties for use with cement. Depending on the charges of the surfactants, it can be anionic, nonionic, or a mixture of both. It is sensitive to temperature; mechanical energy; and, because of the charges of the particles, to free ions that promote coagulation. A high concentration of anionic surfactant can be added to the dispersion to enhance its chemical and physical stability. As cement slurry containing latex polymer-dispersion hydrates, the polymer particles coalesce and bond strongly to cement particles and other solids with which they come into contact. Latex cement is one of the best materials for cementing lateral liners set through windows in multilateral wells.
APPLIC ATION

Experimental Procedure. Conventional gas-migration-control cement and latex cement were tested to determine the better cement for multilateral completions. Thickening time, transition time, fluid-loss, rheology, and compressive strength were tested as specified in American Petroleum Inst. Spec. 10. Flexural strength was tested with the three-point loading method per American Soc. for Testing and Materials Standard D790. Samples were molded to size, 0.5 in. wide, 5.0 in. long, and 0.125 in. thick. The samples were supported in a horizontal position at a span of 4.0 in. A load was applied to the sample at a continuous rate at the center of the span until yield or brittle failure occurred. Strengths were calculated from the linear portion of the load/deformation curve. Permeability, tensile strength, drying shrinkage, and shear bonding were not tested. Results and Discussion. For latex cement, most properties improved as the concentration of latex dispersion was increased. Improvement was particularly significant for fluid-loss control and flexural strength. Compressive strengths were actually reduced with increasing latex-dispersion concentrations. Slurry rheology improved moderately as more latex dispersion was added. Transition time, a property important for gas-migration control, and thickening time decreased as latex concentrations increased. When compared with a comparable nonlatex gas-migration-control slurry, the latex-cement properties were far better

In June 1996, Sarawak Shell Sdn. Bhd./Sabah Shell Petroleum Co. Ltd. drilled and completed the first multilateral well with a cemented junction in the AsiaPacific. The well was a dual-lateral well with two 7-in. liners. The first liner was set conventionally through the 95/8-in. casing shoe, while the second was set from a window milled in the 95/8-in. casing. After cementing the second liner, the liner stub of the second liner had to be cut and retrieved to allow the first liner to be completed. The 7-in. liner was set on bottom through a window measuring 18 ft with 23 ft of it inside the 95/8-in. casing. One aluminum centralizer was installed 13 ft above the top of the window. The slurry used contained 2.0 gal/sack of latex dispersion. An additional anionic surfactant was added to stabilize the slurry. The latex-cement slurry was pumped downhole at a rate of 6 bbl/min with an average pressure of 750 psi. A cement-bond log was run 48 hours later and indicated that the bond was excellent throughout the liner section. Approximately 72 hours later, the liner stub was cut with a rotary mill shoe and retrieved with the guidestock. The cement sheath around the liner stub was still intact and strongly bonded to it. The cut section was smooth, with few cracks. The well was successfully completed and is producing at the expected production rate.

In multilateral wells, liners are set from a main wellbore, exiting through windows milled in the main casing string. The liners are then cemented, and the cemented sections left in the main wellbore (liner stubs) are cut and retrieved later. The liners are cut with a rotary mill shoe that cuts the liners
AUGUST 1997

This article is a synopsis of paper SPE 37623, Cementing Multilateral Well with Latex Cement, by R AbdulRahman, SPE, BJ Oilwell Services Sdn. Bhd., and A. Chong, SPE, Sarawak Shell Sdn. Bhd./Sabah Shell Petroleum Co. Ltd., originally presented at the 1997 SPE/IADC Drilling Conference held in Amsterdam, 46 March. Please read the full-length paper for additional detail, illustrations, and references. The paper from which the synopsis has been taken has not been peer reviewed.
849

C E M E N T I N G

O P E R A T I O N S

WELL CEMENTS THAT REDUCE RATE OF PENETRATION

In Saudi Arabia, old wells are being reentered and completed horizontally to maximize oil recovery. Short-radius wells are being planned for these applications. The initial sidetracking is taking place in the caprock (anhydrite) directly above the Arab-D. BACKGROUND The goal of this study was to develop a cement that would have an ROP less than that of the zone to be sidetracked. The more resistant the cement is to drilling, the better chance of successfully steering the drillstring. Ideally, a high-strength cement is placed in a lower-strength formation. The anhydrite caprock directly above the producing zone, where the sidetracking was to occur was not considered a lower-strength formation. Most cement laboratories have equipment that meets American Petroleum Inst. Spec. 10. This laboratory equipment is used to determine physical properties of cement slurries and set cements. More realistic data can be obtained when job procedures, sequences, and operations are more closely simulated. Drilling rate of penetration (ROP) is one parameter that is used by drilling-rig personnel to determine cement hardness. In the laboratory, the compressive strength is used to estimate the resistance to drilling. Usually, a highercompressive-strength cement has a lower ROP . Laboratory tests are usually not performed on samples of cement used for plugging a well. ROP SIMUL ATOR DESIGN CONSIDERATIONS Functional Design. Laboratory simulation of field conditions is never perfect. The ROP device must fit on a benchtop and must be somewhat similar to a drilling rig. The laboratory drill press should have the same ROP and should rotate at the same speed as the actual drilling rig. Weight on bit (WOB) cannot be simulated. The drillbit design is required to be similar to that of a drill bit used in the field. Drill-Bit Rotational Rate. The standard rotational speed for drilling the anhydrite
850

section is 80 to 100 rev/min. Initially, a standard coring-type drill press fitted with a single-phase, 110-V, 60-cycle/sec motor rated at 0.5 hp and 1,725 rev/min was tested. The motor powered the drill bit through a belt-pulley mechanism that allowed the rotational drilling speed to be varied from 470 to 1,950 rev/min by changing the beltpulley configuration. Because this is greater than the actual 80 to 100 rev/min that will be used to drill the sidetrack, the 110-V single-phase motor was replaced with a threephase, 230-V , 60-cycle/sec motor with a variable-speed motor controller. This motor was rated at 0.75 hp and 3,450 rev/min. The motor controller allowed the drilling rotational speed to be reduced to 80 to 100 rev/min. WOB. Field weight-on-bit values could not be simulated easily in the laboratory. A 9in.-radius pulley was attached to the drillbit lower shaft. An 18-lbm weight was hung over the pulley sets to distribute constant weight on the bit. Normal-water-ratio cement, 118 lbm/ft3, was drilled with this device at 6.4 ft/hr, which is comparable to drilling rig ROPs. Drill-Bit Design. A practical yet economical drill bit was designed for this project. A cone bit was too large for use in the laboratory, so a diamond coring bit was used for all drilling applications in this study. Drilling Fluid Selection. Numerous anhydrite samples were cut into 0.250.251.0in. rectangular prisms. The samples were weighed and then placed in stainless-steel screen cages. The cages containing the samples were placed in different water-based muds and hot rolled for 16 hours at 235F and 500 psi. The samples were removed, and their weight loss recorded. The mud that dissolved the least amount of the anhydrite sample was selected to drill all cement and core samples. ROP AND CEMENT DENSITY It is generally thought that the ROP is inversely proportional to the cement density. Heavier cements have a higher solids-towater ratio and generally a higher compres-

sive strength. Reducing the water and adding dispersants to the cement resulted in an approximately 20% reduction in ROP . For cement slurries with densities greater than 136 lbm/ft3, hematite was added to achieve the higher densities. ROP was reduced to approximately 4 ft/hr in cement with ground hematite. EVALUATION OF MATERIALS It was thought that harder substances in the cement matrix would lower ROP and help divert the drillstring out of the original borehole and into the formation. The set cement had to bond strongly to the inert material so that drilling the test specimens would cut these harder materials instead of removing them from the cement matrix. Ceramic fracturing proppants were used for the harder material because they are generally very hard and are readily available to oilfield operations. Equal concentrations of four different ceramic proppants were mixed with a standard cement slurry. The materials that produced the two lowest ROPs were selected for further evaluation. To determine the effect of proppant concentrations, the concentration of ceramic proppant was varied. As proppant concentration increased (i.e., density increased), ROP decreased. However, the ROP values were not below the target ROP of the core samples. Another study was conducted to determine the effect of particle size of the inert material in the cement matrix on ROP . Proppant sand of 40/60, 20/40, 10/20, and 8/12 mesh were used in this study. The results of this study indicate that the larger the particles in the cement, the greater the reduction of ROP . This article is a synopsis of paper SPE 37821 Well Cements That Reduce ROP (Rate of Penetration), by S.S. Jennings, SPE, Saudi Aramco, originally presented at the 1997 SPE Middle East Oil Show and Conference held in Bahrain,1518 March. Please read the full-length paper for additional detail, illustrations, and references. The paper from which the synopsis has been taken has not been peer reviewed.
AUGUST 1997

C E M E N T I N G

O P E R A T I O N S

NEW SOFTWARE DETERMINES OPTIMUM PLACEMENT OF CENTRALIZERS

A powerful computer program, CentraPro Plus, has been developed. It uses a mathematical simulation model to calculate the optimum spacing of centralizers to obtain the best standoff at a given borehole location. It runs under the Windows operating system and has been installed in more than 100 locations worldwide.
CENTRALIZERS

Centralizers are used to optimize the hydraulic flow during the cementation process and to minimize drag forces caused by differential sticking. During the cementation process, the mud used to drill the well is replaced by cement. If the pipe is not centered, pockets of mud may remain on the narrow side of the annulus, preventing uniform cementation around the pipe. There are two types of centralizers, bow type and rigid. Bow-type centralizers are made of two collars with a number of metal springs attached to them. Because the characteristics of the metal springs are known, the deflection of a centralizer under a given load can be calculated and the distribution of open area between the pipe and the borehole wall can be calculated.
COMPUTER PROGRAM

urements of measured depth, inclination, and azimuth. Directional survey data can be imported by use of a modified commaseparated-values format. A well file contains all relevant data on each casing or liner string. Each string can be divided into different sections to simulate the real well. Parameters, such as centralizer type and spacing, hole size, friction factor, and pipe size, can be varied for each section.
DATA PRESENTATION

Results of the simulation are presented in a way that can be easily read and interpreted. Screen displays include graphs of well profile, depth vs. torque, and depth vs. hook load as well as a tabular presentation of the input and output. A complete report of standoff, required number of centralizers, and stop collars, as well as the values of torque and drag are printed in tabular and graphical form. The tabular data can be exported as an ASCII file, so results can be post-processed in other applications.
C ALCUL ATIONS

sis indicates whether the liner or casing can be run to bottom. Drag is the result of the normal force and the friction factor. Differential sticking of a pipe section not centered properly can be modeled by increasing the friction factor. Rotational torque is based on the normal force vector and the friction factor between the inner side of the centralizer collar and the outer radius of the pipe. By summing the torque starting from the shoe up to the rotary table, it can be determined whether a string can be rotated and whether a connection at a particular location in a string can withstand the rotation without being overtorqued. The model used is a static model that does not account for dynamic peak forces resulting from torsional pipe oscillations.
DISTRIBUTION PHILOSOPHY

An intuitive user interface allows even the most inexperienced users to enter the required data quickly. A spreadsheet calculation system with input/output cells enables the advanced user to change the default data and perform customized calculations. The user can customize the program to use the units of measure and provide output format of their choice. These choices are stored as defaults for future use. Centralizer data and pipe information are stored in databases. The centralizer database contains such data as identifier, geometry, and additional information needed to select an appropriate centralizer for a specific application. The pipe database contains approximately 1,000 combinations of size, weight, and type. In addition to tubing and casing, it contains data for regular and heavy-wall drillpipe and drillcollars. A well-path description is necessary for the calculations. This can be defined with directional survey data containing meas AUGUST 1997

Equations contained in American Petroleum Inst. (API) Spec. 10D were revised and improved. The centralizer deflection is calculated with the results of the restoring force test, performed according to API Spec. 10D. The restoring force is defined as the capability of a spring-bow centralizer to push the pipe radially away from the borehole wall. The restoring force causes a friction force on the borehole wall. The normal force vector is the perpendicular force on a pipe section or a centralizer at a particular position. It is the combined force of the vertical weight component of the buoyed pipe and the tension component perpendicular to the borehole axis. The friction coefficient between the centralizer and the pipe or the formation is used to calculate the drag force and the resistance to rotation. The friction factor depends on the type of mud and on the centralizer/pipe material pair. Standoff is the distance between the pipe and borehole wall when the pipe is fully centered and the actual minimum distance under normal force and sag conditions. Drag-force analy-

Because understanding the origin and quality of the input data and knowing how to interpret the output of the program is important to achieving optimum centralizer placement, this program is offered as a service to customers. The program is not distributed outside of Weatherford Oil Tool. Personnel who use this program are specially trained in program operation and data interpretation. Distribution is controlled with an individual serial number assigned to each copy. This ensures that the same version is used worldwide.

This article is a synopsis of paper SPE 38130, Planning the Cementing Job Incorporates Data Management and Technical ExpertiseA New Software to Calculate the Optimum Placement of Mechanical Cementing Products, by Holger Kinzel, SPE, and Thomas Koithan, SPE, Weatherford Oil Tool GmbH, originally presented at the 1997 SPE Petroleum Computer Conference held in Dallas, 911 June. Please read the full-length paper for additional detail, illustrations, and references. The paper from which the synopsis has been taken has not been peer reviewed.
851

C E M E N T I N G

O P E R A T I O N S

IMPROVED PERFORMANCE OF LIGHTWEIGHT CEMENT SLURRIES

A new engineering approach to lightweight-cement slurries has resulted in systems that can successfully isolate long sections of openhole. The 10.5- to 13-lbm/gal slurries prevent losses into weak formations and can eliminate the need for lead/tail slurries and multistage cement jobs. High early-compressive strengths reduce waiting-on-cement time, can withstand perforating, and are resistant to brine and acid attack. Replacing excess mix water in the slurries with common granular materials is the key to the new technology. To cement successfully across and isolate formations with low fracture gradients requires the use of lightweight slurries. Conventional lightweight slurries usually have a high water-to-cement ratio that results in (1) long waiting periods during cementing, (2) slow development of compressive strength, and (3) a relatively permeable cement sheath that is subject to acid and brine attack. However, a new engineering approach has essentially eliminated these lessdesirable characteristics in lightweight cements. The resulting high-performance lightweight slurries reduce the rig time, logistics, and costs associated with conventional slurry designs. Desired slurry and set-cement properties have been obtained by optimizing the particle-size distribution of a dry-blend cement. The new slurries have properties that are as good as 15.8-lbm/gal cement, including compressive strengths in excess of 2,000 psi at 10.5 lbm/gal (the minimum value accepted by industry standards) and 4,000 psi at 12 lbm/gal, as well as good response to perforating. The set cements show exceedingly low permeability (1 d) and reduced porosity, providing effective resistance to acid and brine attack. The new lightweight cements have demonstrated good performance when used in low-fracture-gradient wells, in geothermal wells, and as a single-stage replacement for two-stage cement jobs (the lead/tail cement is eliminated).
DESIGN METHOD

slurry. In this approach, the solid is a dryblend cement, while the porosity of the slurry is the volume of liquid divided by the total volume of liquid and solid. One of the major changes introduced by this method is that the specific gravity of the dry blend is systematically deducted from the required slurry density for a given porosity. Thus, an appropriate dry blend must be chosen first. The preparation of the dry blend is based on concrete technology. Indeed, as for a concrete, a mixture of a number of granular classes of materials is selected. To select the dry-blend specific gravity required for a given slurry density, materials of different specific gravities and sizes have to be identified. They are simple products: oilwell cement, silica flour, and lightweight minerals. The main parameter for assessing the quality of the granular mixture is its water demand, which is the minimum quantity of water that must be added to the powder to obtain mixability. Optimizing the particlesize distribution and the ratios between the compounds in the dry blend enables the design of liquid concreteslightweight

slurries with low porosity, good mixability, and good rheology. Optimized dry blends have a specific gravity ranging from 1.55 (for 10.5-lbm/gal slurry) to 2 (for 13lbm/gal slurry). The broad range of densities that can be covered shows clearly the design flexibility arising from this new method. Apart from good mixability and rheology, these slurries also are stable and can undergo reduced thickening times at temperatures as low as 100F . The upper static temperature considered for this technology is 450F .
SET-CEMENT PROPERTIES

Several benefits derive from the reduced slurry porosities made possible by dryblend optimization. When set, these slurries provide the same or better mechanical properties as conventional 15.8lbm/gal cement. Compressive Strength. Compressive strength was measured with an ultrasonic cement analyzer while applying 3,000 psi of pressure. Tests were run at bottomhole circulating temperature (BHCT) to corre-

TABLE 1TYPICAL COMPRESSIVE STRENGTH RESULTS WITH VARIOUS LIGHTWEIGHT-CEMENT FORMULATIONS BASED ON OPTIMIZED BLEND Slurry 1 Density, lbm/gal Porosity, % BHCT, F 50 psi BHCT, hours:minutes 500 psi BHCT hours:minutes Compressive strength in 24 hr, psi 10.5 0.5 120 3:53 6:30 2,050 Slurry 2 11.1 0.42 175 7:12 8:12 3,800 Slurry 3 12 0.44 140 6:41 7:33 3,700 Slurry 4 12.5 0.45 230 7:43 8:33 3,200

TABLE 2PERMEABILITY AND POROSITY MEASUREMENTS ON VARIOUS SET CEMENTS Slurry 1* Density, lbm/gal Slurry porosity, % Permeability, d Effective porosity, %
*High performance.

Slurry 2* 12.0 0.45 5-11 22-23

Slurry 3* 15.8 0.58 120-100 32-34

12.1 0.4 8-10 21-22

Slurry density is a function of the specific gravity of the solid and the porosity of the
852

AUGUST 1997

C E M E N T I N G

O P E R A T I O N S

TABLE 3EVOLUTION OF COMPRESSIVE STRENGTH AT 350F BETWEEN 1 WEEK AND 3 MONTHS IN VARIOUS HIGH-PERFORMANCE, LIGHTWEIGHT CEMENTS Cement 1 Density, ppg Compressive strength, psi After 1 week After 3 months 11.73 3,480 3,550 Cement 2 11.73 3,350 3,350 Cement 3 12.00 4,000 4,220

TABLE 4ACID-RESISTANCE RESULTS FOR HIGH-PERFORMANCE, LIGHTWEIGHT CEMENTS AND REFERENCE 15.8-lbm/gal CEMENTS Cement 1 Density, lbm/gal Latex, gal/sack Weight loss, % In 1 hour In 4 hours 12.0 1.0 4.3 Reference 1 15.8 3.7 21.0 Cement 2 12.0 1.25 0.3 0.9 Reference 2 15.8 1.6 2.8 7.7

Compressive Strengths. Lightweight blends containing silica flour were used to prepare three formulations. For each formulation, cubes were prepared and cured at 300F for 1 week and 3 months, respectively. At the end of the curing periods, the cubes were crushed and compressive strength measured. Table 3 results indicate that the high-performance, lightweight formulations maintain the compressive strength over a period of 3 months. No strength retrogression was observed. Mechanical Behavior. Numerical simulations showed that the main parameter to assess the mechanical durability of a cement sheet is the ratio between tensile strength, T, and Youngs modulus, E: the higher the ratio, the better the behavior. Lab measurements of both parameters on suitable cement prisms indicated that T/E is slightly higher in a optimized lightweight cement than in a conventional 16.4-lbm/gal cement. Resistance to Acid. Laboratory tests compared the acid resistance of high-performance, lightweight formulations with a reference 15.8-lbm/gal cement. Comparisons were done on formulations with and without latex because it has been reported previously cement damage in acid treatment might be prevented by adding latex to the cement slurry. Cubes were cured for 48 hours at 185F and immersed for periods of 1 and 4 hours in mud acid (concentration: 12% HCl/3% HF) at 185F . The cubes were weighed before and after the acid-testing period. Table 4 indicates excellent behavior from the high-performance, lightweight cement (cement 1) compared with the 15.8-lbm/gal cement without latex. The weight loss is minimum (4.3 %) in the lightweight cement vs. a weight loss of 21% in the 15.8-lbm/gal slurry. Better acid resistance is observed in the systems when latex is incorporated into the slurries. However, in the 15.8-lbm/gal cement, weight loss is approximately 7% vs. a value lower than 1% in high-performance, lightweight cement.

spond to a large temperature difference between the bottomhole static temperature and the static temperature at the top of cement (sometimes encountered for long cement columns), high geothermal gradients, or both. The results listed in Table 1 (time to reach 50 psi, 500 psi, and compressive strength at 24 hours) demonstrate that the onset of compressive-strength development is not compromised in such systems. Moreover, the development of early compressive strength is very fast, as indicated by the relatively short elapsed time between 50 and 500 psi. This quick development of compressive strength results in reduced rig time waiting on cement. When the plateau is reached, the process evolves slowly and some compressive strength is still gained. The values of final compressive strength are always very high in such systems, varying from 2,000 psi at 10.5 lbm/gal to more than 3,000 psi at greater than 11 lbm/gal. Permeability. Permeability measurements were conducted on various lightweight formulations with the optimized blends. Results were compared with those obtained with a reference 15.8-lbm/gal slurry. Cubes were prepared and cured for 7 days at 185F and 3,000 psi pressure. At the end of the curing period, suitable cores were prepared for air permeability measurement by applying 3,000 psi overburden pressure. Two measurements were taken for each system. Table 2 results indicate that high-performance, lightweight systems exhibit low permeabilities compared with conventional cements.
AUGUST 1997

Effective Porosity. With the same systems as for permeability measurements, effective porosity was measured on cores by use of a mercury porosimetry apparatus (final pressure of 20,000 psi). Two measurements were run for each system. Table 2 results show that effective porosity is reduced in the high-performance, lightweight systems; porosity is about 21 to 23% in such systems and 32 to 34% in conventional 15.8lbm/gal cement. Perforations. Several tests were carried out at room temperature and atmospheric pressure to simulate perforations in a high-performance, lightweight system (12 lbm/gal). A laboratory-scale model simulating borehole geometry was built, and the same slurry was poured into the annulus. The 3-fthigh model consisted of an inner steel casing (4.5-in.-diameter) surrounded by a 7in. length of polyvinyl chloride tubing that simulated the interface with a formation and an outer 9-in. steel casing that was filled with water before the test to isolate the system. Concurrently, cubes made from the slurry were cured at room temperature/atmospheric pressure and crushed every day to monitor the development of compressive strengths. Once the cement had developed compressive strengths in excess of 2,000 psi, subsequent perforation tests were carried out with different charges. Several 0.4and 0.56-in. charges were arranged helically on a gun inserted inside the casing of the model simulating the borehole geometry. At the end of the perforation test, the cement sheet had been demolded and no shattering of the cement was observed.

This paper was taken from an original manuscript submitted by E. Moulin and P . Revil, Schlumberger Dowell. The manuscript has not been peer reviewed. No reprints of this article are available.
853