IADC/SPE 62745 Cements for Long Term Isolation Design Optimization by Computer Modelling and Prediction

Gino di Lullo SPE, Phil Rae SPE, BJ Services Company

Copyright 2000, IADC/SPE Asia Pacific Drilling Technology This paper was prepared for presentation at the 2000 IADC/SPE Asia Pacific Drilling Technology held in Kuala Lumpur, Malaysia, 1113 September 2000. This paper was selected for presentation by an IADC/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 International Association of Drilling Contractors or 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 IADC or SPE, their officers, or members. Papers presented at the IADC/SPE meetings are subject to publication review by Editorial Committees of the IADC and SPE. 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 Today, development of high early compressive strength (CS) is a key performance specification for oilwell cementing. Early strength is important to ensure structural support to the casing and hydraulic/mechanical isolation of downhole intervals. Delays in strength development cause significant amounts of lost time due to the need to Wait-On-Cement. Thus, drilling operations cannot proceed and the rig must sit idle until the cement is deemed hard enough to continue. The actual development of strength in cement systems is dependent on a number of factors. The type of cementitious material is important due in part to chemistry and in part to granulometry. Slurry density is considered critical lower density has, traditionally, been associated with lower strength. Temperature is a key parameter and, to a lesser extent, the pressure. Less appreciated, is the influence of the many types of additives that are included in slurry formulations. Correct selection of cement and additives allows slurry tailoring to achieve "a strong cement" that will support mechanical stresses that happen during the well productive life. Since cement is the primary means of isolation, its integrity as a sealing material should be paramount. Under confining pressure, cement compressive strength is several times higher that the values obtained from an unconfined API test. In fact, the ultimate compressive strength of a wide variety of slurries is similar to that of the surrounding reservoir rock, regardless of initial slurry density. In the long term, more important than compressive strength is the ductility/compliance that set cement has to support downhole stresses imposed by pressure and temperature oscillations during drilling, work-over and production operations. In the

vast majority of well operations, cement failure occurs under tensile loading, not under compression. Poor appreciation of this has led to the routine use of higher density slurries for many applications. Such slurries may be not only unnecessary but also positively harmful, since they can produce more brittle cements that fail under downhole stresses. This paper describes a method of designing cement slurries using a simulator that models cement setting and strength development. Instead of obsolete rules of thumb, mechanical properties of set cement, formation and casing responses to cyclic stresses during the well life are used to optimise designs for early times and long term. Finally, a few case histories are presented, demonstrating the benefits of this new design methodology to produce cost effective cement slurries for one isolation. Introduction: The most important objective of any primary cement job is to provide isolation from producing zones up to the surface and this should be accomplished over the entire life of the well. No fluid movement, either gas or liquid should be possible at any time through the cemented annulus. In the annulus there are three possible paths for fluid movement; the interfaces between cement/rock and cement/casing and the cement matrix. Poor mud removal is normally identified as the major source of communication problems, although poor bonding at the interface can occur even when mud cake or oil films have been completely removed. Mud removal is not included in the scope of this work and it is assumed that state of the art techniques are used on every well to eliminate this problem. Cement bonding can also be affected by slurry properties like fluid loss and free water. However, cement adherence to the formation and casing is primarily affected by cement shrinkage and by stress changes induced by downhole variations of pressure and temperature. These occur mainly inside the casing but can also originate in the formation. By convention, when hydrostatics allow, two types of slurries are used to cement a casing string. A neat (or tail) slurry, with a density ranging from 15.5 lbm/gal to 16.5 lbm/gal depending on cement type and well BHST (bottom hole static temperature), is used to anchor the casing. The second type is a filler slurry that has lower density and,

GINO DI LULLO AND PHIL RAE

IADC/SPE 62745

generally, less demanding performance specifications. It develops lower strength and has higher permeability than the tail slurry but is still adequate to protect and support the casing. Both, the anchor slurry and the filler slurry are used to reduce the overall cost of a well. Several rules of thumb have evolved to simplify slurry design; unfortunately some of them are based on old technology and outdated price/performance models that are no longer relevant today. A good example relates to cement compressive strength, its relationship with slurry density and the mechanical properties of hard cement. It is not clear why it is still normal practice to overpressure the formation, inducing damage or risking lost circulation problems, with cement slurries that are 2-8 lbm/gal heavier than the mud that was used to safely drill the well. Actually, under confinement, all cements eventually have a strength essentially approaching that of the surrounding formation, regardless of the initial slurry density. Unfortunately, correlations drawn between sonic (compressional) wave attenuation (CBL) and compressive strength have led to the misguided conclusion that the higher the strength, the better the bonding. This correlation between attenuation and compressive strength is quite poor; a much more respectable correlation exists between compressional wave attenuation and acoustic impedance (2) but this is rarely used due to general unfamiliarity in the industry. In addition, the strength/attenuation correlation was based on unconfined compressive strength measurements and does not reflect down hole confined conditions, where cement becomes ductile instead of brittle and compressive strengths are 2-9 times higher (Figure 1). Note that Young's Modulus, under different confining pressure, remains practically unchanged. The CBL actually measures the coupling at the interfaces, the amount of volume filled and the porosity of the set cement. It is obvious that the porosity is related to the solids content of a set cement but not necessarily to the cement content. Further complicating matters, is the fact that other factors such as effective coupling and gas, water or mud channels can affect compressional wave velocities and their attenuation. In the field, this has led to the common practice of trying to overcome poor bonding by increasing slurry density, instead of analysing and understanding the fundamental reasons for bond failure. Neat cement slurries (no additives) mixed as per API specifications, will produce brittle, high strength (in general, Young Modulus > 1.3E6 psi and Poissons ratios < 0.20) and low permeability cements (k < 0.001 m.d.). However, cement coupling, or hydraulic bond to the formation and casing, are not at all related to compressive strength. Cement bonding is not accomplished by a chemical reaction but by a physical interaction. Therefore, bonding is more related to cement ductility and casing roughness. It is true that cement adherence, or shear bond, is somehow related to cement compressive strength through its tensile strength relationship but the addition of other materials to the cement slurry can greatly affect the result.

Anchoring and Supporting the Casing To minimize rig time, and hence well cost, there is a need to speed up cement setting and bonding at the interfaces. This ensures that the rig doesnt stay idle for long periods waiting on cement to set. This is, obviously, more important offshore and on deep wells where rig costs could be an order of magnitude higher than for average onshore wells. High axial loads are imposed on the casing string and surrounding cements by landing or suspension methods and later by drilling or other operations. This is especially true if the well is highly deviated, horizontal or completed using multi-lateral systems where the cement sheath may be subject to shattering and loss of bond eg. due to mechanical shock or temperature and pressure cycles. Casing support can be measured in the laboratory using the shear bond test or the tensile strength test and calculated by the following formula (1): F =9.69 x Ts x D x H, Where: F = force or load to break cement bond, pounds Ts = tensile strength, pound per square inch D = outside diameter of casing, inches H = height of cement column, feet The force, or load, to break the cement, in the worst case scenario, is the weight of the casing plus the weight of the complete drilling string while tagging or drilling the shoe. Table 1 shows that only 15-psi tensile strength is required to continue drilling, using this formula. In this evaluation, we used the heaviest casing for its size with only 5% equivalent cement fill up of tail slurry (based on casing length) to support the casing and no additional support from the filler slurry. For neat cement slurries, the tensile strength can be safely estimated by dividing the cement compressive strength by 10. Thus, in the above scenario, only 150-psi compressive strength is required; in other words, less than one third of the accepted industry standard (500 psi). However, in lighter or extended slurries, the ten percentile ratio (tensile: compressive) does not apply and, depending on the slurry composition, the ratio could be as high as 1:2. Also, the thickening time and strength development can be accelerated with proper additives. Infact over 95% of the slurries in use today in the field will develop more than 150 psi, in less than 5 hours after placement. Figure 2 is a UCA plot comparing compressive strength development of three cements (13.3 lbm/gal slurries) at identical down-hole conditions; the slurry formulations are included in Table 2. Note how both money and time could be saved using a properly designed lightweight slurry and not waiting for 500psi strength. Therefore, for anchoring and casing support, we can reduce the cost and simplify the operation by using a single low-density slurry. A properly designed slurry could have a density of close to 12.5 lbm/gal or, depending on the hydrostatic control pressure required, 0.5 to 1 lbm/gal heavier than the drilling mud will suffice. Table 3 contains the slurries and their strength properties for comparison purposes.

IADC/SPE 62745

CEMENTS FOR LONG TERM ISOLATION - DESIGN OPTIMIZATION BY COMPUTER MODELLING AND PREDICTION

Zonal Isolation at the Interfaces Zonal isolation is mainly dependent on the hydraulic bond of cement to the interfaces and the permeability and integrity of the cement sheath. For a good seal in the annulus, the cement must be flexible enough to behave as a packing element does in a packer. During the setting process the cement undergoes volumetric shrinkage (2, 3) that can be as high as six percent. This can affect hydraulic bonding and the permeability of the cement sheath. Two different types of cement shrinkage have been recognized; bulk or outer shrinkage and inner or chemical shrinkage. Bulk shrinkage affects bonding at the interfaces and varies depending on slurry composition. Increasing the solids content or adding magnesium and/or aluminum salts can reduce, or even eliminate, bulk shrinkage. These salts can actually induce expansion of the concrete under unconfined conditions but they can also increase the final cost of the slurry. Under confinement, cement becomes ductile and it deforms easily under pressure. Every time there is casing contraction, due to hydrostatic pressure or temperature reductions in the well bore, the cement will elastically expand until it becomes plastic and the hydraulic bond is broken (formation of micro annulus). Improving the elasticity of cement or its flexural and tensile strength is an elegant solution to prevent debonding. This can be accomplished by adding materials (3, 4, 10) such as latex, polyvinyl alcohol (PVA) and/or semi-inert mineral fibers. Flexural strength can easily be measured in the laboratory or estimated for neat slurries using 0.25:1 (flexural strength:compressive strength) ratio. In the case of lighter slurries, however, particularly those containing mineral fibres, this ratio can be as high as 1.25:1. In other words, such lightweight slurries exhibit greater flexibility and can provide better zonal isolation. Zonal Isolation in the Cement Matrix There are two possibilities for flow through the body of the cement; (a) High cement permeability that could be related to an inner shrinkage of cement or high porosity cement due to poor slurry design and (b) Macroscopic cement sheath failure due to longitudinal and radial cracking under pressure or temperature cycles (8, 10). Inner shrinkage has to do with the reduction of the absolute volume (or increase in porosity), that occurs when water and cement combine to form hydrates. This process, which starts just after mixing cement, continues until cement hardening has been completed weeks, months or even years later. The chemical shrinkage rate is low or non-existent during the hydration period, very high during setting and initial hardening and very low from then onwards during the final hardening. Therefore, in the short term, during cement setting,

highly mobile fluids (gas) can invade the cement matrix and percolate in the well, producing channels that will greatly affect sonic wave (CBL) response and possibly affect zone isolation. Slurries with low free water and fluid loss, right angle set (short transition time between liquid and solid) or those containing specific gas control additives can reduce or prevent this problem (5, 6). Also, the addition of semi-inert micro size particles, (preferably flexible fibers) of appropriate granulometry will help to control fluid movement through the cement matrix. In the long term, radial cracking of the cement sheath can happen at any time during the life of the well. Radial cracking is normally due to tangential and flexural forces rather than compressional forces. These forces start at the annulus boundaries, when the casing or plastic formations (salt domes) expand (11). They can also occur due to compressional forces in extreme cases, such as in an annular gap between a soft and a hard casing or a soft casing and a hard formation. At the limit, under very high loads (variations in hydrostatic pressures or well temperature), compressional forces could destroy the cement sheath by compaction of matrix porosity. Therefore, to improve zonal isolation, (while maintaining set cement porosity and permeability as low as possible) we need to increase ductility (tensile and flexural strength, 9) of the set cement. Increasing slurry density (to reduce porosity) is not the answer, however as it will increase compressive strength and induce brittleness. Instead, permeability and ductility improvements can be accomplished with lighter slurries and additions of low density semi-inert materials, such as micro-size mineral fibers, that reduce set cement porosity and improve the flexural and tensile properties of set cement (reference 8). In addition when fluid loss properties are required in the slurry, vinyl base additives or latex should be used instead of HEC or polyamine derivatives. Mathematical Simulator Even though Flexural and Tensile strength measurements are commonly used in the construction industry and testing equipment is inexpensive and commercially available, they are not common in oil field applications. Yet it is clear from field results and the published literature that these parameters are far more important from a zonal isolation perspective than compressive strength, per se. In order to promote a better and more widespread understanding of this fact, a "cement strength simulator" was developed. This software, based on a series of empirical equations, predicts, with good accuracy, the development of cement mechanical properties and the response to downhole stresses. As such, the program can be used for preliminary design purposes to assess the applicability of any given cement slurry to accomodate the stresses imposed on it during well completion and production. The empirical relationships take the form of a series of polynomial equations programmed using fuzzy logic, to take into account synergistic and/or stochastic effects. The algorithm produces a chart similar to that produced by an Ultrasonic Cement Analyser (UCA) with cement strength

GINO DI LULLO AND PHIL RAE

IADC/SPE 62745

evolution versus time. Cement strength evolution exhibits an S shape curve (Figure 2 & 4), with three distinct phases. The first phase represents the dormant period during which early hydration is proceeding. The cement has developed no significant mechanical strength at this time. This phase is deemed to be complete when the slurry has attained 5 psi compressive strength. This transition point is designated as the Kick-Off Time(KOT). The second phase is typified by a rapid gain in mechanical strength (high slope) and represents the "early hardening". Finally, the third phase describes the decelerating reactivity (low slope) that signifies "final hardening". Each of these phases is described by a separate equation or series of equations containing terms that relate to the physical properties of the slurry, the chemical interactions of the components and the synergistic relationships between them, as a function of time and temperature. The final strength is calculated at 48 hours since, at this time, over 90 % of the total strength will typically have developed in oil field cements. Also, this is the minimum time recommended before running bond logs to evaluate zonal isolation. Both, the KOT and the compressive strength are calculated, within certain confidence limits, based on results obtained with several cements around the world. Thus the final compressive strength development plot is a band instead of a single line. Figure 5 is a simulation of slurry 5 (UCA in Figure 4) listed in Table 1. The accuracy of the simulator is within the accepted testing error limits for this type of test as defined by ASTM, or better than 75%. Figure 3 is a plot of actual UCA compressive strengths versus software predictions, for comparison purposes. The Flexural and Tensile Strength are calculated based on the calculated compressive strength at 48 hours. Again, there are several terms in the polynomial equation that take into account the effects of individual additives or the synergistic behaviour exhibited by combinations of certain additives. Table 3, compares actual and predicted properties for all 5 slurries listed in table 3. The mechanical properties of cement (Young's Modulus and Poissons ratio) can either be input or estimated by the simulator. The estimation is done by empirical correlations. The Poissons ratio is correlated with the flexural strength and Young Modulus is correlated to the compressive strength. The simulator calculates the changes in the tangential and radial components of stress in a cement sheath due to changes in wellbore pressure, far-field total stress, and wellbore temperature. These induced stresses are calculated by solving the differential equation governing the radial component of the displacement field in the casing, cement, and formation. The solution of the displacement field has 6 constants that are determined from 6 boundary conditions that impose continuity of the radial components of stress and displacement at the inner surface of the casing, the casing-cement and cementformation boundary, and far-field. As a result, the six constants are functions of the inner and outer diameters of the casing, well bore diameter, and the elastic constants (Young's Modulus and Poisson's ratio) of the casing, cement, and formation.

The calculation follows the procedures outlined by Thiercelin et. al. (11). The following assumptions were made to simplify the calculation: 1. Casing, cement, and formation possess linear elastic, isotropic properties 2. Isotropy of horizontal far field stress components 3. Plane-strain solution (i.e. no change in height of cement) 4. No theta-dependence (in cylindrical coordinate system) of the stress or displacement fields. Although the radial and tangential stress components are calculated in the casing, cement, and formation as a function of the distance from the well bore axis, only the maximum values at the interfaces are used to calculate cement sheath failure. This simulator calculates changes in the stress components in cement due to changes in well bore temperature. Simplifying assumptions were made to allow a closed-form solution of the radial component of the displacement field. It is assumed that the casing is uniformly heated and expands into the cement before the cement has time to absorb heat. Thus, temperature gradients are neglected and temperature change is used to calculate the maximum stresses. The results have good agreement with the more exact numerical solutions. The formation-cement boundary is also neglected (which is equivalent to assuming that the elastic constants of the cement and formation are identical), because temperature effects are almost instantaneous at the cement casing interface, before heat is dissipated through the cement sheath. The solution of the displacement field in the casing and cement has four constants. The constants are determined from four boundary conditions that impose continuity of the radial components of stress and displacement at the inner surface of the casing, the casing-cement boundary and far field. In this case the change in horizontal stress is assumed equal to zero. As a result, the four constants (elastic constants for casing and cement, and thermal expansion coefficient of the casing) are functions of the inner and outer diameters of the casing. Radial and tangential components of stress are calculated as a function of the radial coordinate. However, only the average values of these stresses in the cement are used to calculate cement sheath failure. This approach seems to provide better agreement with the maximum induced stresses that are derived from the more exact numerical solutions of this coupled thermal-elastic problem. The simulator also provides an estimate of bonding and isolation length. The theoretical CBL amplitude is calculated for a dual receiver tool and the barrier length required for zonal isolation, for any specific case, is calculated based on gas pressure and formation flow capacity, a similar approach to those described by Rae et. al. (12) and Economides (13). These parameters are estimated using flexural and tensile properties of the set cement in addition to the compressive strength. Finally, the cost of the slurries is also part of the output to assist engineers in the optimization of cost-effective

IADC/SPE 62745

CEMENTS FOR LONG TERM ISOLATION - DESIGN OPTIMIZATION BY COMPUTER MODELLING AND PREDICTION

designs. Using the simulator To illustrate the use of the simulator and its application, we have selected a shallow well where the last 2 strings cemented were a 9 5/8 casing and a 7 liner, as per the schematic in Figure 6. All five slurries in Table 2 were tested as a single slurry for cementing the liner, based on future operations on the well. With regard to zonal isolation, all of them passed the minimum required length adjacent to, or just above, the zone of interest. The worst case scenario was using the bentonite slurry that required a minimum of 75 ft (Table 3), which is 3 times smaller than the shale barrier height. After the liner is cemented in place, the top of the liner is normally tested, (see Table 4). Applying a pressure through the casing or though tubing set in the 9 5/8 casing, will make the 7 liner expand due to a ballooning effect; in this case there is no change in temperature. All five cements will be damaged in front of the unconsolidated sandstone. However, as the cement in front of the barriers remained intact, it is possible to test the liner to 2000 psi. The gas zone will be acid fracture stimulated in this well. Thus, during the job there will be an increase in pressure and a decrease in temperature due to cool down effects (see Table 4). If the job was performed through casing or through tubing with a packer set at the top of the liner, the only slurries that could be used are numbers 1 & 5. Notice that slurry number 1 has a density of 13.3 lbm/gal but contains micro-fibers, which provides it with the necessary properties. The neat slurry and the two lightweight slurries will have cement sheath failure in adjacent zones compromising zonal isolation. Also, notice that the bentonite slurry is the only one that failed under compression. Finally, the well will start producing gas, creating a pressure drop inside the casing and a temperature increase in the upper zones. In this scenario only the 15.6 lbm/gal slurries could be used, as all the extended slurries exhibit failure in adjacent zones. As a cost-effective alternative we used slurry number 1 but mixed at 13.6 lbm/gal with a PVA fluid loss additive instead of CMHEC and it passed all three simulations. The simulator is not only intended for use in preliminary design but also in field applications to help explain mysterious phenomena. Examples might include cases where the cement has magically disappeared from behind the casing after an excellent job with no reported problems. More routinely, the programme can be used to educate engineers on the complex interaction of the forces that are involved in achieving and maintaining zonal isolation. A better understanding of these will help to optimize slurry designs and reduce the cost of effectively cementing a well. Case Histories: Venezuela: This well is situated in a difficult area to drill in Eastern Venezuela. The wells are quite deep (15,000 ft) and hot (350F), with several zones prone to lost circulation

throughout the complete interval but especially at 5,500 feet. Also, there is presence of CO2 and H2S in almost all pay zones, including the shallow ones. The development of this field is being reactivated to target dry and associated gas in the deeper zones. Three casings and two liners are used to construct these wells. The casing setting depths are: 20 at 700 feet, 13 3/8 at 3,500 feet, 9 5/8 at 8,500 feet and the 7 5/8 and 5 liners are set at 13,800 feet and 15,000 feet, respectively. The 9 5/8 casing is the most difficult to drill and cement as there are five low pressure gas zones (consolidated sandstone) intercalated with active shale that are prone to lost circulation from 5,000 to 6,000 feet. In addition, there are two highpressure gas zones at around 7,000 feet. Perfect zonal isolation in this interval is mandatory and cement sheath integrity is expected to last over the entire life of the well. Formate mud with density ranging from 10.0 to 11.5 lbm/gal is used for drilling this section, dictating the use of lightweight slurries. Also higher static temperatures encountered in the deeper intervals (up to 350F) mandate the use of cement retrogression additives, even though the static temperature in this interval is only 210F. The cement in this interval is subjected to severe cycles of temperature and pressure stresses while drilling the next intervals. An increase in hydrostatic pressure of 2.5 lbm/gal (equivalent to 1250 psi) and an increase in temperature of over 60F, while drilling the liners indicated the use of ductile cement to prevent cement sheath failure. The Cement Simulator program was used to help on the pre-design of these slurries. The main parameters studied were mechanical response to cyclic stresses and gas migration prevention while drilling the liners. The simulator suggested the use of slurries with at least 670-psi compressive strength and flexural strength better than 310-psi. Compressive strength, flexural and tensile strength, gas flow and the traditional fluid loss, free water, thickening time and rheology / compatibility tests were carried out to design the slurries and verify the simulator predictions. Mud removal and displacement were accomplished with 100 bbls of a water base pre-flush (to eliminate compatibility problems) and 100 bbls of a multi purpose cement spacer, both of them designed for turbulent flow. Two slurries were used, a tail slurry covering all the lost circulation zones up to 4,900 feet and a lead slurry up to 200 feet inside the 13 3/8 casing. The final slurry designs used were: Lead Slurry: Class H + 30% hollow ceramic spheres + 3% natural mineral fibers + 3% activated charcoal fibers + 0.5% anhydrous sodium silicate + 1.5 gps AMPS fluid loss additive + 0.8 gps colloidal carbon dispersion + 0.01gps anti-foam. Density : 1.0 lbm/gal (11.8 at downhole) Fluid Loss : 21 cc in 30 min Free Water : 0 ml Thickening Time : 5:50 hours,

GINO DI LULLO AND PHIL RAE

IADC/SPE 62745

Compressive Strength : 48 hours: 1,250 psi Flexural Strength : 48 hours: 269 psi. The Lead slurry was tested for: a. Increase in hydrostatic pressure of 2,000 psi; the simulator estimated a minimum CS of 454 psi and a minimum FS of 140 psi were required for cement sheath integrity. b. Increase in temperature of 50F; the simulator estimated a minimum CS of 93 psi and a minimum FS of 162 psi were required for cement sheath integrity. c. A combined increase in pressure and temperature (1,000 psi and 30F respectively); the simulator estimated a minimum CS of 283 psi and a minimum FS of 166 psi were required for cement sheath integrity. Tail Slurry: Class H + 25% silica flour + 10% hollow ceramic spheres + 1.2 % PVA bonding additive + 0.45% CMHEC fluid loss additive + 0.3% dispersing agent + 1.9% anhydrous sodium silicate + 1.8 gps colloidal carbon dispersion + 0.01gps anti-foam. Density : 12.64 lbm/gal (13 at downhole) Fluid Loss : 19 cc in 30 min Free Water: 0 ml Thickening Time : 3:34 hours Compressive Strength : 48 hours: 1,380 psi Flexural Strength : 48 hours: 334 psi. The Tail slurry was tested for: a. Increase in hydrostatic pressure of 2,500 psi; the simulator estimated a minimum CS of 622 psi and a minimum FS of 222 psi were required for cement sheath integrity. b. Increase in temperature of 60F; the simulator estimated a minimum CS of 138 psi and a minimum FS of 240 psi were required for cement sheath integrity. c. A combined increase in pressure and temperature (1,000 PSI and 50F respectively), the simulator estimated a minimum CS of 428 psi and a minimum FS of 312 psi were required for cement sheath integrity. The simulator predicted values for these 2 slurries for TT, CS and FS were within the 25% accepted error for this type of testing. Also, the CBL amplitude from the log was below 12 mV (except for the zones where loss circulation ocurred), as predicted by the simulator. Trinidad: Trinidad off-shore - The 13 3/8-in. surface-casing primary cement jobs were executed for over a year utilizing a silicate extended lead slurry. Recently, a study using the simulator was done in order to come up with a more cost effective slurry. Cementing Parameters 30-in. conductor driven to + 675 ft, 13 3/8-in. casing depth at + 4,727 ft.

Openhole size 16 in. Openhole excess 70%. BHST120F / BHCT101F Previous Lead Slurry: Class G + 0.65 gps sodium silicate extender + 0.30 gps dispersing agent + 0.25 gps synthetic retarder + 0.25 gps PVA based fluid loss additive and bonding agent + 0.05 gps anti-foam + sea water. While trying to maintain the slurry performance, we tried various slurries in the simulator to improve early compressive strength development and reduce WOC. This was the reason for the addition of an aluminum salt. The New design: Class G + 0.60 gps multi purpose cement additive + 0.18 gps sodium aluminate extender + 0.05 gps anti-foam agent + sea water. The new designs, as shown, are simpler, have better properties and have been used without any problems. This has reduced the cost by almost 15% and produced enough strength to actually eliminate the need for a tail slurry without altering the 12 hours WOC requirement. Offshore Malaysia: To counter problems arising from the concomitant presence of shallow gas and lost circulation in a challenging offshore environment, a new, high performance cement system was developed. The properties of this system were examined using the simulator and optimized accordingly. This new system has a density of 14.0 lbm/gal, which is between the commonly used 12.6 lbm/gal lead and 15.8 lbm/gal tail slurries. Because of its moderate density, it generally requires less cement additives in order to be gas-tight for shallow gas control purposes. Coupling this fact with the utilization of highly efficient cement additives results in a more effective and economical shallow gas control cement slurry. Apart from resisting gas migration, the lightweight system develops adequate compressive strength for this application. Typical 12-hr compressive strengths vary from 300-800 psi, depending on temperature. The strength values at 24 hours are normally in the range of 800 1,500 psi, easily competent for support and isolation. This type of slurry also has a flexural: compressive strength ratio higher than 0.32 and the isolation length required for 0.4 psi/ft gas gradient is < 25 ft. Such performance eliminates the need to use a standard density tail slurry and helps improve cement job execution. Using only a single slurry simplifies the cementing operations because of easier mixing and better continuity. Switching slurry densities (ie. from lead to tail) can be problematic and, statistically, a significant number of cement job problems are associated with this transition period. Continuity during cement jobs is important because, for example, a variation in lightweight slurry density of only 0.3 lbm/gal may affect both compressive strength development and slurry viscosity values by more than 30 percent. Thus, the resulting slurry meets all specifications, making it ideal for cementing across intervals where problems of weak formations and shallow gas co-exist. Similar slurries could

IADC/SPE 62745

CEMENTS FOR LONG TERM ISOLATION - DESIGN OPTIMIZATION BY COMPUTER MODELLING AND PREDICTION

also be used more widely in many circumstances where, today, conventional lead and tail slurries are used. Adopting this approach would greatly simplify the mixing procedure, as noted above, allowing better continuity and greater control over that most important of slurry properties density. Furthermore, slurries of this type offer substantial improvements in drill-out time while still easily exceeding any compressive strength limits for casing support and structural competence. The final point worth noting is the long term safety, security and environmental integrity of the well. Most oil and gas wells suffer leakage over their lifespan indeed most regulatory bodies accept, and expect, a certain gas leak rate from wells. Incorporation of an active gas absorbent in well cementing slurries offers the tantalizing possibility of reducing such leak rates by an order of magnitude. This could reduce the risk of escape of hydrocarbons to surface, or interzonal communication subsurface, thereby improving wellsite safety, environmental stewardship and simplifying well abandonment. Up to late 1999, seventeen wells were successfully cemented with this cement system in Malaysia. There have been no cases of annular pressure build-up in any of these wells. Conclusions: 1. A computer simulator that models the evolution of cement mechanical properties has been developed. This tool can be used to help understand the complex processes involved in achieving and maintaining well integrity and zonal isolation. 2. High cement compressive strength is not crucial for cement sheath integrity and zonal isolation and should not be used as the basis for designing cement jobs. Mechanical properties of set cement, under cyclic downhole stresses, should be used to guarantee well integrity. 3. Less that 150 psi compressive strength is required for anchoring and supporting casing. Elimination of the tail slurry is possible by using slurry densities higher than 12.5 ppg and/or 1 ppg heavier than the drilling mud, reducing the cost and simplifing the operation. 4. Casing expansion created by higher internal casing pressures or temperature increases can create radial cracks in the set cement in the annulus, which cause loss of annular zonal isolation. 5. Casing contraction, caused by reductions in hydrostatic pressure or lower temperatures, can result in the formation of a microannulus with attendant loss of annular isolation, particularly to highly mobile fluids (i.e. gas). 6. Low compressive strength cements are more ductile than other cements and are better suited to withstand stress cycling. Addition of certain polymers (PVA, StyreneButadiene latex) or, preferably, natural mineral fibres can also be use to enhance cement flexibility. 7. Generally, the addition of natural mineral fibers allows the use of low-density cement slurry to cement the entire well, eliminating the need for tail slurry, or multistage cementing.

Acknowledgment: The authors wish to thanks the management of BJ Services for the permission to publish this paper. Also, many thanks to all those in the laboratories and the field who helped support this development, especially Russell Maharidge and Norberto Briggiler. References: 1. George O. Suman, JR. and Richard C. Ellis: World Oil Cementing Hand Book, 1977 2. J.J. Jutten, P.A. Parcevaux and D.D. Guillot, Schlumberger: Relationship between Cement Composition, Mechanical Properties and Cement Bond Log Output. SPE 16652, Dallas, TX, September 27, 1987. 3. P.A Parcevaux and P.H. Sault, Schlumberger: Cement Shrinkage and Elasticity: Anew Approach for a good Zonal Isolation. SPE 13176, Dallas, TX, 1984 4. Dan Mueller, BJ Services: An Evaluation of well cements for use in High Stress Environments. JPT , 1999 5. Gino Di Lullo and James Tan, BJ Services: An Evaluation of Gas Control Additives. OSEA 92167, Singapore, 1,December, 1992 6. J.C. Hibbeler and Gino Di Lullo - BJ Services and Michael Tay Shell-Malaysia: Cost-effective Gas Control: A case Study of Surfactant Cement. SPE 25323, Singapore, 8, December, 1993 7. K.R. Backe, O.B. Lile, S.K. Lyomov, Harald Elvebakk and Pal Skalle, Norwegian U. OF Science and Technology: Characterizing Curing-Cement Slurries BY Permeability, Tensile Strength and Shrinkage SPE Drill & Completions 14, 3 September, 1999. 8. K.J. Goodwin, Mobil E&P Services Inc. and R.J. Crook, Halliburton Services: Cement Sheath Stress Failure SPE 20453, December 1992 9. R. Abdul Rachman, Phil Rae, BJ Services: A Simpler, More Effective Method FOR Cementing Surface Casings in Shallow Gas Zones: A Case History SPE 54286, Jakarta, Indonesia, 20 April 1999. 10. R.B Carpenter, J.L. Brady and G.G. Blount, Arco Oil & Gas Co.: Effects of Temperature and Cement Admixes on Bond Strength SPE 22063, Anchorage, May 31 1991. 11. M.J. Thiercelin, J.F. Baret, Schlumberger and W.J. Rodrigues, Intevep S.A. Cement Design Based on Cement Mechanical Response SPE 38598, San Antonio, TX, 8 October 1997. 12. P. Rae, BJ Services and D. Wilkins, Marathon Oil Company and D. Free Dowell Schlumberger: A new Approach TO THE Prediction of Gas Flow After Cementing SPE 18622, New Orleans, Louisiana, 3 March 1989. 13. Michael J. Economides, Dowell Schlumberger: Implications of Cementing on Well Performance, Well Cementing Book, Chapter 1, 1987

DI LULLO G., RAE P.

IADC/SPE 62745

Table 1
Casing diameter Inches 20 13 3/8 9 5/8 7 Casing Length Feet 1000 7000 15000 22000 Casing Weight per Feet 133 72 54 32 Tail Cemented Height feet 50 350 750 1100 Casing Weight Pounds 133000 504000 802500 704000 Drill String Weight Pounds 15500 108500 232500 341000 Total Weight Pounds 148500 612500 1035000 1045000 Tensile Strength Required for CSG Support PSI 14 11 11 9 Tensile Strength Required for CSG + Drilling String Support, PSI 15 14 15 14

Table 2
Slurry Type API Cement Type Multi Purpose Additive 15% 50% 15% Extender Bentonite 6% 0.25% Additives Extender Sodium Silicate 0.4% 0.36gps F. Loss Additive CMHEC 0.45% 0.45% 0.2% 0.15% 0.80%* Cement Dispersant SNSC 0.25% 0.2% 0.15% 0.45% 13.3 13.3 13.3 15.6 15.6 13.6 Density ppg Thickening Time h:m BHST F

1 2 3 4 5 6

A A A A A A

4:26 4:16 3:44 3:30 3:15 3:18

135 135 135 135 135 135

* PVA Fluid Loss Additive

Table 3
Slurry Type TT Min. 4:26 4:16 3:44 3:30 3:15 3:18 CS 24H PSI 1710 900 800 2450 1650 2020 Actual Test Data CS 48H PSI 1950 1150 1060 3250 2050 2280 FS 48H PSI 365 296 313 776 581 458 TS 48H PSI 166 128 170 355 275 208 KOT Min. 4:51 4:29 2:54 3:32 3:23 2:40 CS 24H PSI 1523 910 875 2338 1739 1809 CS 48H PSI 1626 1119 1094 2960 2076 1894 FS 48H PSI 440 291 289 770 659 603 Simulator Predictions TS 48H PSI 195 123 121 326 260 249 CBL mV 6.4 8.8 8.8 3.8 3.4 5.4 Required Isolation Height, ft 59.0 67.6 74.7 9.4 8.2 40.3 Young Modulus MPSI 0.49 0.44 0.43 0.95 0.74 0.66 Poissons Ratio 0.21 0.18 0.19 0.18 0.24 0.22

1 2 3 4 5 6

IADC/SPE 62745

CEMENT FOR LONG TERM ISOLATION - DESIGN OPTIMIZATION BY COMPUTER MODELLING AND PREDICTION

Table 4
Evaluation at Formation
Top Shale

Operation
Test Liner Hanger Test Liner Hanger Test Liner Hanger Test Liner Hanger Test Liner Hanger Test Liner Hanger

Delta Pressure
25000 25000 25000 25000 25000 25000

Delta Tempe.
0 0 0 0 0 0

Cement Sheath Integrity

Slurry 1 OK Fail by Tension OK OK OK OK OK Fail by Tension OK OK Fail by Tension
Fail by Tension & Compression

Slurry 2 OK Fail by Tension OK OK OK OK OK Fail by Tension Fail by Tension Fail by Tension Fail by Tension
Fail by Tension & Compression

Slurry 3 OK Fail by Tension OK OK OK OK OK Fail by Tension Fail by Tension Fail by Tension Fail by Tension
Fail by Tension & Compression

Slurry 4 OK Fail by Tension OK Fail by Tension OK OK Fail by Tension Fail by Tension Fail by Tension Fail by Tension Fail by Tension Fail by Tension OK

Slurry 5 OK Fail by Tension OK OK OK OK OK Fail by Tension OK OK OK OK

Slurry 6 OK Fail by Tension OK OK OK OK OK OK OK OK OK OK

Unconsolidated Sandstone Medium Shale Sandstone Bottom Shale Limestone

Top Shale

Fracture Job Gas Zone Fracture Job Gas Zone Fracture Job Gas Zone Fracture Job Gas Zone Fracture Job Gas Zone Fracture Job Gas Zone Bottom Zone Gas Production Bottom Zone Gas Production Bottom Zone Gas Production Bottom Zone Gas Production Bottom Zone Gas Production Bottom Zone Gas Production

3600 3700 3800 3850 3950 4000

-70 -70 -75 -75 -85 -90

Unconsolidated Sandstone Medium Shale Consolidated Sandstone Bottom Shale Limestone

Top Shale

-550

20

Fail by Tension
OK

Fail by Tension
OK

OK

OK

OK

Unconsolidated Sandstone Medium Shale

-575

15

OK

OK

OK

OK

-600

10

Fail by Tension Fail by Tension

Fail by Tension Fail by Tension

Fail by Tension Fail by Tension

OK

OK

OK

Sandstone

-625

OK

OK

OK

Bottom Shale

-650

Fail by Tension

Fail by Tension

Fail by Tension

OK

OK

OK

Limestone

-675

Fail by Tension

Fail by Tension

Fail by Tension

OK

OK

OK

10

DI LULLO G., RAE P.

IADC/SPE 62745

Table 5
Property Slurry wt (ppg) Slurry yield (cu ft/sx) Mix Water (gal/sx) Thickening time Free Water (mLs @ 101F) Fluid loss (@ 1,000 psi & 101F) Comp. Strength @ 120 F & 12 hrs Comp. Strength @ 120 F & 24 hrs Rheological properties (600, 300, 200, 100 RPM) Previous Design 12.6 2.15 11.92 6 hrs 0 175 100 450 45,25,20,13 Simulator Prediction 12.6 2.15 12.55 8.00 hrs API 173 570 Actual Design 12.6 2.15 12.54 5 hrs 0 135 500 750 115,73,53,32 Simulator Prediction 12.6 2.15 12.55 5.49 hrs API 357 698 -

IADC/SPE 62745

CEMENT FOR LONG TERM ISOLATION - DESIGN OPTIMIZATION BY COMPUTER MODELLING AND PREDICTION

11

Figure 1

Young's Modulus vs. Strength

24-48 hour set time 4

3.5

Unconfined Compressive Strength

Confined Compressive Strenght

Young's Modulus (Mpsi) 2.5

1.5

0.5

0 0 2000 4000 6000 8000 10000 12000 14000 16000 Compressive Strength (psi)

Figure 2

12

DI LULLO G., RAE P.

IADC/SPE 62745

Figure 3

Actual CS vs Simulators Min &Max CS

5450
Simulator Min CS Max CS

4450

Simulator Range, PSI

3450

2450

1450

450 450

1450

2450

3450

4450

5450

Actual Compressive Strength (UCA), PSI

IADC/SPE 62745

CEMENT FOR LONG TERM ISOLATION - DESIGN OPTIMIZATION BY COMPUTER MODELLING AND PREDICTION

13

Figure 4

Figure 5

2.5E+03

160

140 2.0E+03 120

Compressive Strength , psi

100 1.5E+03

Temperature, F

80

1.0E+03 60

40 5.0E+02 20

0.0E+00 0 4 8 12 16 20 24

Time, hours

14

DI LULLO G., RAE P.

IADC/SPE 62745

Figure 6

9 5/8 Casing, N-80

4500 feet Top Shale 4700 feet Unconsolidated Sandstone 4850 feet Silt-Stone 5150 feet 5250 feet 7 Casing, N-80 Consolidated Sandstone Oil Zone Bottom Shale 5450 feet 5,500 feet Limestone, Gas Zone Gas Gradient 0.5 PSI/ft

BHST: 135 F Temperature Gradient 1.0 PSI/100 ft Fracture Initiation Gradient: 1.10 PSI/ft Mud Density: 10.6 PPG