SPEAADC 39279 Gas Migration Eliminated Through Correct Cement Design Including Elastomers S Talabani, SPE, Wester Atlas international Logging Services, G. Hareland, SPE, New Mexico Tech, and MR. Islam, SPE, UAE University Abstract thas been proven through laboratory experiments that the three types of gas migration through a cemented annulus can be eliminated by designing the correct ‘cement mixture. The first type of void and, therefore, {gas migration can occur between the casing and the cement. By adding the correct amounts of magnetite to the cement, this void, and therefore the possible cause of ‘gas migration, can be eliminated. The second possible type of void generation is between the cement and the borehole wall where the filter cake formed at the borehole wall adversely affects the bonding process. By using a special material, Anchorage Clay, this bonding, can be improved to the extent that the gas migration between the borehole and the cement can be eliminated. Te third and most complicated process is the pressure changes appearing in the cement during the setting phase. A double sinus wave pressure response during this setting time generates fractures in the microstructure of the cement. The correct amount of water as well as retarders is crucial for the best results during the dehydration process of the cement. By adding the correct type of elastomers, this pressure variation during, the setting of the cement can be eliminated. Elastomers are known to counter-react the pressure behavior during, the setting process, This eliminates the pressure 263 variations and, therefore, the micro cracks, The three above mentioned effects are strong functions of temperature and pressure, and the cement design for a well would have to be carefully planned since a well has both a temperature and pressure gradient with depth. ‘This paper discusses the individual components necessary for gas-leak climination and gives a ‘quantitative field example where all the correct additive ‘volumes have been designed as a function of depth. The paper gives a clear guideline for designing the total climination of gas migration during a cement job. In addition, this paper clearly addresses all the gas ‘migration problems related to cementing operations. Introduction Cement slurry passes through two cycles of building its structure during the gelation period’. Figure | shows schematic of the duration of the first and second cycles of the total gelation time, In the first cycle, the cement slury builds itself with time in a three-dimensional structure (thixotropic behavior)’. However, in some local areas, this structure collapses, releasing some trapped water’, This situation takes place mainly in the presence of tight formations. When cement faces porous formation the preflush cement loses part of its ‘water to the formation as a spurt filtrate. This results in an incomplete cement-water reaction ‘The cement cake in the spurt loss area is the weakest part of the cement since it did not receive enough water to complete the reaction. The micro- cracks will first appear in the spurt loss area even ‘though filtrate control materials are used with the cement. ‘Temperature and pressure are the two physical properties that contribute to the final cement seting ‘The effect of temperature has the most impact on the interface between the casing and the cement, Similarly, pressure affects the setting of the cement slurry that undergoes a phase transition from liquid to solid. ‘The cement is not compatible with the casing metal, thus the2 ‘S. TALABANI, G. HARELAND, MIR. ISLAM SPE 29279 attachment between the two materials is weak. This ‘weak attachment can be affected by both pressure and temperature variations. In addition to pressure and temperature, there ‘are many parameters in cement jobs that cannot be controlled or their effects totally eliminated”, eg. an eccentered casing that causes non-uniform slurry flow, the control of free water of the slurry under borehole conditions’. Experimental Analysis Dozens of experiments were conducted using Class G cement under wellbore conditions. Results show the following 1. tis very important to know that the addition of any additive (except inert solids) causes the change in the cement reaction equation. Thus, itis essential thatthe least numbers of additives are used. 2. Experiments show that the addition of filtrate loss control materials should be low enough not to create ‘thin film between cement and casing, but high enough to prevent any spurt loss in porous zones ‘These laboratory experiments are important to conduct before any cement job in the ficld Knowing porosity and permeability of some key zones in which preflush might occur is important. A few experiments should be conducted under the pressure and temperature conditions prevalent in the formation in order to observe how much water is lost before and after the filtrate material agents are added. Extra volume of water should then be added to the slurry by the amount of the filtrate loss in order to compensate for this loss. This amount of ‘water is needed to complete cement-water reactions. 3. Extra water of 20-30% by volume (of extra water added to wet polymer powder) needs to be added to compensate for the polymers that trap additional water. For example, if a biopolymer is added as a filtrate control agent, one must be aware that some of the biopolymers ‘have long molecules with hydrophilic end that locally trap some water and prevent this water to react with cement, Hence, if x is the volume of water needed to compensate for the addition of the polymer powder, the real volume should be 1.2 to 13x in order to compensate for this entrapment Coment Stages The cementing process goes through three main stages ‘until cement is set. They are: 264 Cement Slurry Stage. For this tage, the cement flows and reacts with the continuous phase simultaneously. This means that he cement slur) at diferent postions in the borehole el have differen theology. For this Stage, cement fllows the power law model for cerain Deiod of time ffom the cement-water mixing prio” After the end of this period, the cement slurry follows ‘the Harschel-Bulkly rheological model, given by: Toy +k a) were, Kis the vscosiyy constant and y is the she me Gelation Stage. This is the most critical stage that cement setting goes through. For this stage, cement slury loses most of its continuous phase (free water) 10 cement reaction in the form of filtrate loss to the formation, Cement in this stage is no longer mobile. It has a 3-D thixotropic shape in which it makes a bond between cement grains. These bonds are not strong enough to bear the load of the pocket above the cement slury. During this stage, cement passes through the first cycle of expansion (from the gain of cement-water reaction), and contraction (complete reaction and molecule connections). At a certain point time during is stage, cement collapses locally and receives the surrounding non-reacted water. This collapse takes place as a result of weak 3-D structures for which the weight of cement above any local area becomes higher than the 3-D Thixotropic structure. During the final part of this stage, the cement rebuilds itself locally after further reactions with the leftover water. For this stage, maintaining pressure on the annulus might help eliminate the local void areas created during the life span of the stage" During this stage, the consistency of the cement slurry increases by more than 150 times from its original value when it is first mixed’. The viscosity gradient of cement, however, remains extremely low because all the particles make stronger 3-D bonds at this stage. Consequently, two cycles emerge for which ‘cement expand and contract twice. ‘Cement Set Stage. This stage follows the end of the gelation stage for the last cement shrinkage ‘occurs after its final reaction with the leftover water. At the beginning of this stage, local cement that faces porous media needs additional water (equal to the amount of the water lost to the formation) to complete its reactions. Because there is no water available, someSPE 36279 cement particles will have close bonding (the ones that hhave enough water to complete their reactions). On the contrary, other particles with incomplete reaction will be removed farther away, creating some void areas within the cement structure. If there is a gas present in the formation, it will flow and fill these voids within the cement structure. ‘After certain time, gas under pressure can penetrate and connect itself through these voids creating a micro- network locally. This local area will be the beginning of the micro cracks within the cement body. The presence of mud can reduce the generation of these micro cracks. ‘Types of Micro Cracks Thete are two types of micro cracks present in oil-well cements, They are: Micro-Annulus. Two types of micro annular spaces can ’be found in oil wells. One is the cement-casing micro annulus and the other one is the cement-formation micro annulus. The first one is known to create micro cracks ‘more often than the second one. This is due tothe high contrast between the metal and the cement structures. The casing metal expands as it receives extra heat during hydration reaction of cement. After certain time the metal cools down and its temperature levels off with the surrounding for which the temperature i dictated by the geothermal gradient. Hence, both cement (at its final stage) and easing shrink ata certain time. Because their coefficients of thermal expansion are not equal, tmicro annular cracks are formed, This situation will nt occur along the casing string since in some local areas ‘cement can reshape and attach itself to the casing under the effect of the weight above it The second type of annular micro cracks mostly occurs at local areas where the formation has a porosity. This is because ofthe mud and/or the cement crack between the formation and the cement. Micro Fractures. This type of micro cracks occurs within the cement body itself. It results from the latest expansion-coniraction cycle during the last part of the cement gelation period. This type of cracks are easily found in the laboratory experiments conducted under reservoir pressure and temperature’, Under reservoir pressure and temperature conditions and for certain flow rates in eccentered annulus, cement can replace most of the mud but cannot enter the mud cracks, ‘The thin film of mud attached to the casing stays in certain intervals where the flow regime is expected to be laminar’ Micro fractures were generated in these intervals after the cement set 265 GAS MIGRATION ELIMINATED THROUGH CORRECT CEMENT DESIGN INCLUDING ELASTOMERS 3 Elimination of Micro Cracks In this section, the remedies to the formation of micro cracks are discussed following different types of micro cracks. Eliminating Micro Annulus Adjacent to the Casing. Because steel metal and cement are two. diferent rmaterils with a wide contrast in their thermal expansion, densities and other properties, departing from each others occur during and after the cement is set. To eliminate this problem, one should select @ material that can be partly embedded in cement while remaining close (0 steel metal property-wise. Ironite Scavenger (magnetite) was the right candidate in Iaboratory experiments. Ironite Sponge creates, magnetic field around casing and causes most of its particles to be concentrated closer to the casing than the rest of the cement body at a high temperature’, The concentration of Ironite Sponge is found to be the highest inthe are close (othe easing string, However, the rest ofthe cement body has very litle of the Irnite imbedded in it. Compressive strength test shows the strongest part of cement to beat the part attached to the casing. An addition of 7-10% by weight of cement was found to be enough to eliminate the micro annulus in 92% of the experiments conducted using. Irnite Sponge’. Reducing Cement-Formation Micro Annulus. This is the most difficult part of the main problem, especially in the enlarged hole sections and permeable zones. In an enlarged hole, mostly laminar flow regime can be achieved. This resulls in annular cement flow leaving. part of mud and/or mud cake behind. This part of mud cor mud cake will prevent cement spurt loss. It also prevents the cement-formation attachment. On the other hand, cement might lose part of its water to the permeable zones, creating a weak cement cake at the formation face. A solution to this problem was by adding Anchorage clay to the cement and inject it as a patch cement facing the zone’. This clay consists of about 50% Ilite, 30% muscovite and 20% chlorite which swells with time as temperature increases. In fact, the clay merely causes more local expansion of cement, thus squeezing the mud cake further. Also, the presence of clay increases the resident capillary pressure significantly, helping the plugging process. The critical amount of this material is 6% by weight of cement above which cement shows a rapid decline in its compressive strength.4 'S. TALABANI, G. HARELAND, MLR. ISLAM SPE 39279, Eliminating Cement-Body Micro Fractures. This problem results from the two cycles of cement expansion - contraction. An ultrasonic equipment can be used on cement sets to find out the extent of micro tures. Both cycles of expansion-contraction are functions of pressure and temperature down hole. Maximum change in the cement volume recorded was 4.6% by weight of cement under room temperature and 100 psi pressure”. The addition of latex liquid to cement did not solve the problem entirely. It further complicated the problem by maintaining the liquid phase under high temperature and later forming its own networks locally. Addition of solid tar extracted from cil was also checked. This solid particles did not dampen the expansion-contraction but it only filled some pores. The only solution to this problem was found to be the addition of certain rubber powder to the cement. Rubber powders have different elasticity and other properties. Certain amount of rubber powder should be added under local conditions of each zone’. Optimum Cement Elastomer Mixture If one wants to achieve zero-change in cement volume, the following equation can be applied for most formations AV ig (t= Avra Foo Fee” Foti J, 7 Of course, in order to achieve zero change in the volume, the above equation must be equal to zero, yielding: tam -_(Satemn) o Vacs Var py Where P, T are the surrounding pressure and temperature, respectively. If the zone contains gas, Pore pressure is used for P. In order to determine relevant compositions of the ‘optimum cement job, following experiments are recommended: First Experiment. The first experiment has the following protocols. © Add all the previous materials to the cement powder, plus certain essential additives need to be added, + Mix cement with water for certain time (20 ‘minutes) and pour it into a closed container of known volume. © Apply zone pressure and temperature to the slurry in order to simulate downhole conditions. ‘© Leave the cement for a certain period of time until it is set © Open the container and measure the exact volume of the cement shrunk. This volume will be the volume of certain rubber powder that needs to be ‘added in the next experiment, Second Experiment, For this set of experiments, there might be a need that more than one experiment are repeated to establish meaningful trends. The protocol of this set of experiments is: * Add the same mixture to the same powder as before. * Mix the cement powder with a volume of rubber powder equal to the amount found in the first experiment. * Conduct the same experiment procedure (as in the first experiment) and apply same pressure and temperature tothe system. * Open the container after the cement is set, and use a syringe to measure the amount of cement shrunk. Repeat the same experiment by adding a volume of rubber powder equal tothe shrunk volume. * After three experiments are conducted, plot the change in volume with the volume of rubber powder added to the cement. Figure 2 shows a sketch of the change in cement volume with the percentage of rubber powder added. ‘Theoretically, this plot should be straight line with 45 degree angle, however, it is experimentally found thatthe point at which the slope changes indicates the amount of rubber powder that needs to be added to the cement to eliminate micro fractures, Design of Coment Slurry There are two different ways to design the cement slurry. They are described below: Multiple Patches Design. If itis practical to design several patches of cement, cach for a zone, then a set of experiments needs to be conducted on each zone under its reservoir conditions (pressure, temperature, porosity, and permeability). Several patches should be made with different rubber concentrations and injected in the field sequentially. Bottom zone patches need less rubber powder concentration because of the high temperature and pressure and hydrostatic head on the cement slurry. As pressure, temperature and hydrostatic head on theSPE 39279 cement slurry decrease, its last expansion-contraction during the gelation period results in a greater change in the shrinking volume. Single Cement Slurry. If only one cement slurry is to be used, a set of laboratory experiments should be conducted in order to determine maximum contraction, ‘and thereby, the right amount of rubber powder to be added to the entire cement as one patch. For example, ‘Table | shows a set of experiments that were conducted under different pressure and temperature conditions. As pressure decreased the last contraction of the cement was 4.14 ce per 420 ce cement sample. This means that no further experiments are needed. The data of this table are then plotted (Fig. 3) and an extrapolation is needed up to the surface pressure in order to find out the ultimate cement shrinkage, Equation 3 is then applied as follows: Vena is the extrapolation change in volume from Fig, 3 at reservoir temperature and atmospheric pressure This is because all the cement pockets that later result in ‘micro-annulus would be under minimum pressure. Vere: is the volume of cement used in the experiment (420 ce in our experiments). Vaaue : ig the volume of rubber powder added to the ‘cement to compensate for the shrinkage volume ‘AVoase is change in the volume of the rubber added to the cement slury under applied pressure and temperature, This should compensate for the last part of | rubber contraction when cement is et. Example of Cement Design If 420ce cement slury is used in each experiment, and the contraction volume was around 17.64 ce, the amount of rubber powder added to the cement found to be 23,809 ce (from extrapolation). Equation 3 is later applied as follows: 1764 x 420 ~ 23809 - alec, rubber volume is needed for the last shrinkage. This means that an addition of 200 - 300% of this volume (2 ce for example) is needed to compensate for the last shrinkage in which the rubber powder will be slightly shrunk by the set cement as follows: 1764 x~2 420 23809 ©) 267 GAS MIGRATION ELIMINATED THROUGH CORRECT CEMENT DESIGN INCLUDING ELASTOMERS, 5 x =3.08ce of rubber is needed in which 2.08 ce will be squeezed all over the 420 cc cement sample under ‘conducted pressure. Hence, the volume of squeezed rubber is equal to 100x3.08/420 = 0.73% It is safe to make this squeezing factor 1-2 %. ‘This procedure can later be applied to the field. To be ‘more accurate, one can do this calculation for every pemeable zone by preparing different slurry patches. This will totally eliminate both micro-annulus and micro-fractures in a cased hole. Conclusions Based on the discussion and presentation of results in this paper, following conclusions can be made: 1. Certain laboratory experiments should be conducted prior to conducting any cement job in the field 2. A guideline for caloulating the volume of elastomer is given. If followed properly, this guidelinene can optimize cementing operations 3. Addition of three new materials to the cement prevents any gas migration through cement, but requires carefl design. Nomenclature Shear Stress Minimum Stress Required to Get Certain Fluids Moving, Consistency Index Shear Rate Power Law Index Pressure Temperature HES as Acknowledgements ‘The authorization to publish this paper by Western ‘Atlas Logging Services is gratefully acknowledged. References 1. Talabani, S. and Hareland, G., “New Cement Additives That Eliminate “Cement Body Permeability”, SPE 029269 paper presented at the Asia Pacific Oil & Gas Conference held in Kuala Lampur, Malaysia, 20-22 March, (1995). 2. Walton, IC. and Bittleston, S., * The Axial Flow of Bingham Plastic in a Narrow Eccentric Annulus”, J Fluid Mech, Vol. 222(1991),e S. TALABAN|, G. HARELAND, MLR ISLAM SPE 39273, 3. Bittleston, S. and Guillot, D., “Mud Removal: Research Improves Traditional Cementing Guidelines”, Oilfield Review, April (1991). 4. Keller, S., Commentary, p. 37, Oilfield Review, April (1991). 5. Marcassa, F., Commentary, p. 43, Oilfield Review, April (1991), 6. Talabani, S., “Experimental Study on Gas Migration in Cemented Annulus", MS Thesis, University of Alaska Fairbanks-Fairbanks, AK, August (1992). 7. Talabani, S. and Hareland, G., “ Expansion Contraction Cycles for Cement Optimized as a Function of Additives”, SPE 29471 paper presented at the Production Operation Symposium held in Oklahoma City, OK, 2-4 April (1995), Table 1 ‘Total cement contraction on different experim Experiments | Applied ] Total contraction ] Applied ‘conducted | pressure | co/420cc sample | Temperature Bi 1 1000 76 OF 2 300) 30 200F 3 200) 34 2007 4 750) 358 200F 3 700) a 3008 6 0) a 206F 7 670) a IF g 0 a IF "*Same composition of cement and additives was used Initial setting Final setting Pressure Time, brs Fig. 1 Schematic of different stages of cement setting av ———_____+ % rubber added Fig, 2 Schematic of the cement volume change with rubber added 45 e 4 Bas i 3 2s Eis 4 Sas ° 500 700 900 1100 Pressure, psi Fig. 3 Total cement contraction for different applied pressures.