Putting a Stop to Gas Channeling Ironically, intrusion of gas into the casing-wellbore annulus is sometimes facilitated by the presence of the ‘cement put there to seal it. Understanding gas intrusion is an evolutionary process that has not yet run its full course. Meanwhile, practical efforts to stop the gas hav Gerard Bol Koninklijke/Shell Exploratie fen Produktie Laboratorium Rijswijk, The Netherlands Hal Grant Chevron Services Co. Houston, Texas, USA Stu Keller Exxon Production Research Co. Houston, Texas, USA Franco Marcassa Total Compagnie Francais des Péroles, Centre Scientifique et Technique, Saint Rémy Les Chevreuse, France Jean de Roziores Saint- iene, France Fox hep in preparation ofthis atl, thanks Roy ‘White and Caro lenin, Chewon Services Co, Hous ten es, USA hems vn Vi Koes [prac en Pol aborain,Rpai Te Netheand snd Alin Ducheee, oa Compare Francais des ces, Corre Siemisque Teche, Sin Rey Les Cheese rane ‘A key aim of cementing a well isto provide ‘durable isolation between different forma: tions and the surface. Opposing this is the intrusion of gas into the cementilled ann lus. This can lead to development of a path- way for formation fluids within the cement sheath, either between zones or to the sur face (above). Both are dangerous and poten- tially costly, ‘Of most concern ate channels that allow gas to come to surface. This can occur either via casing strings cemented to surface or in liners where gas emerges at the liner hanger and is then circulated out of the well. The severity of annular gas low ranges fom hazardous fo marginal: rom a blowout toa few psi of residual gas atthe wellhead Interzonal communication is difficult 10 detect but can impair gas production or anced dozens of potential solutions. A gas channel ‘produced in the Iaboratories of Exxon Production Rosoarch Co. show. {ng how gas can create signiticant Channels through coment. recharge an upper depleted zone, possibly affecting production and/or drilling condi tions from adjacent wells. In the extreme, the problem can lead to abandonment of the well. More frequently remedial measures like squeeze cementing ate attempted until gas intrusion is reduced to levels that comply with oil company and regulatory authority standards. Such squeezes are extremely difficult to perform and often unsuccessful, First, the gas chan: rel must be detected and then cement has to be placed into it, often into very small aps. Pressure exerted during the cement squeeze can break down cementcasing, bonds creating new problems, Clearly, there is strong motivation for performing the job conectly inthe first place.‘Coming to Grips with the Problem The chief diffculty in understanding gas rusion is that the slurry evolves with time. A liquid immediately after placement, It pro- _gresses through a series of different states before hardening, This transition, when ‘combined with the complex downhole con- ditions, makes laboratory simulation dificult (ee “From Test Cell to Well,” page 38). Despite this, three separate routes by which gas comes to surface are recognized. All three may be present inthe same annulus. The fist route for gas is via columns of ‘mud (stringers) left in the set cement as a result of poor mud removal (befow). As the ‘mud gels, dehydrates and shrinks, formation fluids can flow through the voids. High pressure Second, gas can flow through microan: uli along the cements interfaces with the ‘casing and the formation (top, right). These ‘microannuli develop because of volume reduction of cement as it sets and cute. Pressure and temperature variations during the life of the well can, at very least, exag: grate the problom, and some experts feel they are actually the main cause. The pres: lence of thick, mushy iter cakes on the for- ‘mation or mud clinging to the casing also increase any microannulus. But it isthe thied mechanism of gas int sion that has most exercised the theoreti- cians and experimentalsts over the past 20, years or so, While the first two mechanisms elfectively bypass the cement sheath, the 36 High pressure {gas zone High pressure [Route two for gas migration: through microannull between either cement and casing formation. ‘or cement and igh pressure gaszone = [Route three for gas migration: the coments micrestructure, called ‘matiix gas channeling. third provides a route within the cement itself (above). in what is called matrix gas channeling, gas. flows through the ‘microstructure of the cement, ot possibly Creates tiny fractures in the setting cement (Gee "Commentary: Stu Keller,” next page). [At first this seems impossible. Ifthe well is killed by the drilling fluid prior to cemen- tation and the slury’s hydrostatic pressure exceeds the formation pressure, then gas should be kept at bay. However, after mud hhas been displaced and cement left static // G08 out fow rate Time, be bbchind the casing, several processes begin that undermine this static state. Fluid loss (dehydration of the liquid phase), sedimentation (etling and packing ‘0f the solid particles) gelation of the slury ‘and shrinkage are all occurring as cement sets. The relative importance of each depends on slurry composition and well conditions which may vary across the inter- val being cemented. Shortly after placement, the still liquid slurry, ceases to transmit its full hydrostatic pressure to the formation. This reduction in hydrostatic pressure has been measured by Exxon Production Research Co, Pressure and temperature sensors were placed on the out- side of the casing, in seven wells. Logging Cable, also clamped to the casing, brought the data to surface (next page, lft). Pressure ‘generally decreased shorly after cement was pumped.” Furthermore, pressure loss was ‘greatest adjacent to permeable zones where fluid loss from the slurry was highest. IF there is a diferential pressure between the slurry and a permeable layer, water can flow from the slurry into the formation. This leads to a volume reduction and, if this reduction is not compensated for, cement pore pressure declines. Compensation becomes harder as time passes, because the ‘cement builds up gel strength a8 a structure How gas enters the annulus. As ‘coment sot, its bility fo transmit Surry empereture pressure, Gas can then enter the cement, Oilfield Reviewis establishing its matrix. Static gel strength is the shear stress required to cause the cement to flow again, and its rate of development depends on slurry composition and down- hole conditions. During gelation there is no significant hycation of the cement grains Eventually, when gel strength exceeds shear stress exerted by the weight of the cement and the mud above it, the cement suppor itself. Once self-supporting, pore pressure within the cement structure is no longer controlled by the hydrostatic load transmitted through the cement. Rather itis controlled by the slurry permeability and shrinkage; this may lead to a substantial decrease in cement pote pressure. Hiycra tion results in an absolute volume reduction fof the cement matrix (shrinkage}—as much a8 6 percent—because the volume of the Cerystalized phases is less than the sum of the volumes of the water and the dry ‘cement particles.» For all these reasons, cement pore pressure ‘can ultimately drop below formation pres- sure (previous page, bottom) and there will bbe room for gas to enter the cement matrix. It gas invades the cement pore spaces, isolation of the formation is jeopardized and the success of the cement job is at rsk:+ coninuedon pag 40 Measuring coment hydrostatic pressure in the field. By tasting seven wells with pprossure sensors, Exxon was able fo show that coment hydrostatic pressure decreares as the coment sats andl’ ‘becomes self-supporting. Commentary: Stu K¢ ‘Thore are generally three types of annular gas flow: channel, microannulus and matrix. Even if you have perfect displacement of mud by the cement, annular gas flow may still occur. By understanding the causos of the prob- om, itis possible to devise approaches that might be beneficial. First, low ld loss slurry is almost univer. sally accepted. Tis cus the volume loss from the ‘slurry since, by definition, a tow fluid loss stury elds alow permeability fiter cake. Reducing the ‘ld loss helps mainain the cement poe pressure atatime wl a tow downward ‘the cement matrix and lower the tendency of gas tofow tough the cement even when there has 2 reduction in cement pore pressure. While the cement matix prme- large, It may inate formation Hed coment that aro responsible for much higher volumes of gas flow. ‘Alhed alternative i to use a slurry that has ‘elayed gel strength development wit right angle sot Delayed gol strength development is ‘advantageous because he tu loss rate fom the ‘cement is a decreasing function of time. This ‘means that when the cement does ge, the pres- ‘ure reduction duet id loss wil be les. Right-angl sti key tobe beneticial because reducing the time the strry spends in transition als ets hor ong the slrry is most yunerabe to 35 intrusion, ‘Beyond modifying the slur, there are physical ‘ways of opposing gas entry. Maintaining pressure ‘on the annulus can help counterae gel strength ment by making the column flow dow ward to compensate for any volume loss at the bottom ofthe column, The ge strength ean gener- ale resistance, so the effectiveness of his m ‘sure depends on the slurry design. Generally we limited by the fear of exceeding the tractre ressuro of the formation and can only exert 2 le of hundred psi. Meverthelss, this i worth in conjunction with ater options, particu larly whore the cement column is relatively short This prompts another techique—Iimiting the ‘eight ofthe coment in areas where annular gas ‘ow is expected. This fs achieved through use of ‘Stu Kale, seo research specie, ting and compltions ‘ivsion, Goan Produc tion Research o, ‘multistage cement jobs and by ting th ‘pumped. Theoretically, use ofa larger annulus ‘an also help cul he pressure reduction asthe ‘cement cluria begins to become selt-suppoting, ‘hough often this fs nt feasible because the cas- ing and hole size are governed by (Quite often with cementing, there ae drawbacks with every potential ation. the factors betor doit ‘Chemical shrinkage has nat been by causing a sigiticant reduction in cement pore pressure, A combination a an oxpansive coment ‘ith some of the other materials discussed might bea good approach. Cement must provide a seal throughout the etre life of a well But fututing ‘and pressure and ai ‘mediate casing ean aso create a miroannulus. ‘Common sense is also needed. If there is 18 pound per gallon mud in the production casing while cement sats inthe annulus, iis not a good [dea from the pot of vow of annuar isolation to thom use diesel ful asthe packer probably want fo use a heavier ld to contraction of the casing that may causo a gap in the annulus, 1: Parceviun PA and Sa PH =Cement Shinkage ae {Benge Nt and Cheng PRA Scale Down ahora ‘ery Tet rere for lain Specie Well Cn ‘Tr The Section of Conant Recipes o Contd Foxton Fi Wigton Ate Crp,” Dobe SPL19522 present atthe oath Pe Anal ‘econ Conerenee and atin San Ani, ‘ows USA, Otobe 11, 1989 2. Cocke CE, Kuck MP ad Madano Red Mes ‘nest Arla rsa ond Temperate Dring Pinay Cemertin” ura of Petoleur Tecweoey $5 ga5Taz0-58 Elsie New Appar Cad Zonal ik. tion poe SPE 13176 presi a the Sh SPE ‘Aa tecnica Conlrece aa Exton, HoUs fom esas USA epee 1619, 1984 4. Sains sey Mand Suton DU: “Tanston Tine ‘Cen Sars tween he Fd an Set Ste, bape SP 9285, presredathe Ssh SOF Arua fethncal Conksence and Eton, Dsl, tens, USA Sepember 2124, 1980. 37From Test Cell to Well ‘There is no American Petroleum Institue (API) recommended procedure for testing cement’ ability to withstand gas intrusion. Furthermore, ‘the large number of variables require anyo interested in testing slurries to make simpliica- tions. Reviewing every method that has been ‘used to lest the gas-blocking capabilities of ‘cement slurries is beyond the scope of his art le. Here are two examples: Rober Brule of Amoco Production Co. and rs have developed a computer-conrolled gas ies a variable hydraulic pres- loss from the slury and estab- lishesa gas pressure dlleratial across the slurry that mitrors formations of citerent pres sures (ght). ‘the heart ofthe apparatus isa moditiod high-omperature/high-ressure AP! tu lass call hat contains the sry. A hollow hydraulic pislon witha 325-mesh scroon is Inserted into ‘the cel through the top. Hydrostatic pressure is ‘simulated by pressurizing te piston with mineral ‘ol (whieh i set pressurized by nitrogen) A Dressure transducer monitors the piston pres- sure. The hollow piston shat is connected by Sterible hose to the top backpressure reg ‘Slurty fitrate from the top ofthe cell pases rough the piston shaft and is collected and ‘measure by electronic balance. ‘Another 325-mesh seren is locatad at th bot- 3 ofthe call balow which there isn source concted toa diferental pressure regu- lator which has is high-pressure sid connect to the bottom sereen, Between this regulator and ‘he bottom screen is second electronic mass ‘owmeter hat measures the simulated formation 425 flow ate int the coll. The low-pressure si othe diferental pressure to the top backpressure regulator And the pres- 38 Mineral oltank valve Pressure regulator Thermocouple — oh | tu eh Bottom fitrate valve Balance Top fitrate ‘al oY valve Top backpressure reguiator Differential pressure Nitrogen gas Diferential Gas fow- Pressure Formation eter regulator valve [Measuring gas intrusion in the laboratory using complex apparatus. Computer con- trol allows Amoco experimenters fo automatically aller variables ike cement hydro- slatle and differential pressure, sure across the diferential pressure regulator is ‘monitored by a transducer control system. Fate trom the bottom ofthe eal is collected through the bottom backpressure regulator and ‘measure using another electronic balance. The pore pressure of the cement sur is measur using two pressure transducers 1 inch 26 mm) and 5 inches [130 mm] trom te boom screen. A ‘thormocouple inserted trough the side ofthe cell ‘measures the cel temperature which Is modite ‘using @ heating jacket. Movement ofthe piston shaft is monitored bya displacement transcueer. ‘The ast assumes a worst case ofa gas zone with enough pormeability to fully invade and pressure charge cemented annulus. After mix: the test slurry i sted for 20 minutes in an atmospheric consistometer at botlomnoteeircu- lating tomperature (BHC) while the tost celts Also proheated to BHCT. Al the same timo, the nitrogen system is readied, With the format {28 Inlet vale closed, a simulated gas formation pressure I applied from gas inlet valve tothe top backpressure valve. The same pressure is also applied tothe backpressure regulator. At this point the slurry Is poured int the top of he cel, ‘and the piston is Inserted and used to apply the intial hydrostatic pressure ‘The gas inlet valve atthe bottom ofthe ‘opened as isthe bottom filtrate valve. This allows slurry flld loss to oecut, The top trate Oifield Reviewvalve remains closed to avoid the formation of a ter cake which coulé inhibit later ow of titrate and gas though the piston shaft. The cel is then heated trom BHCT to bottomboe state tamper {ure using a predetermined schedule that mirors what really happens toa fui once pumping stop. Th hydraulic pressure ofthe piston above the strry simulating the hyrostatic head is radualy roduc ‘Cement pore pressure is continuously monk tored asi falls, and once it approaches the co stant simulated gas formation pressure, the pi ton hydraulic pressure is noted and kept constant {or the rest ofthe tat, At this point, te top f- ‘rae valve is opened, and diferental pressure is applied across the coment lurty from the bottom 1 the eel to the top backpressure regulator. I 25 flow occurs during the est, the low rate is ‘measured bythe mass flowmeter, (Crucial to the relevance ofthe results are the hnyrostate prossure reduction program (which in the test determines the rate o alll ofthe slurry ore pressure, the diferential gas pressure applied across the slury and how these values are sealed down o benchtop proportions. The dectne of coment slrry pore pressure with ime is complex and not fully understood. function of (among other tis evelopment, fd loss and shrinkage tion to downhole temperature and pressure. In {is apparatus most ofthe fall in pore pressure is atibuted to gel strength development, Usi soparate apparatus, the slurry gel strength devel ‘opment versus time is measured under downtole temperature and pressure. From these results, a program for reducing the hydrosiatc pressure transmission that occurs as the slury becomes sol supporting Is caleulated and translated into pressures applied by the piston during the test, ‘Deviing what pressure ditterential to apply across the slurry isnot as straighforward as it ‘may first appear. itis not enough to apply a pres sure gradient calculate by subtracting the pres- April 1991 ‘sure a weak zane in the wll rom that ofthe high pressure zone and dviding iby the distance between the two zones; gas compressibility has tobe considered, For instance, while ina wel, 2 128 zoe at 7,000 pst may be Mowing toward a ‘zone a 6,500 psi. The laboratory simulation will have a gas source at 400 psi flowing towards a lower discharge pressure. To calculate ths dis charge pressure, Boirute apples Darcy's Law for linea flow and assumes thatthe slur bulk por ‘meabiliy an gas viseositis are the same forthe well andthe call. ln this way, low rate per unit ‘ar in tho wol is calculated and then uso to produce the discharge pressure for he cel, Koninklike/Shell E&P Laboratorium has opted {0 use U-tube gas suction apparatus (below). The U-tube sited with cement slurry and gas pres ‘sure applied tothe top of one othe tube legs. ‘The coment i allowed to sot and the gas pressure ‘monitored continuously. Cement permeability o ‘the gas can then be measured by increasing the pressure dlferential across the U-tube by up to 3,000 psi and measuring the resultant low rats. To separate the effect of tow thraugh 2 ‘mieroannuus (ue to shrinkage ofthe coment) Presse ‘wansducor 2,000 psinitiogen pressure soures coment ‘and low through coment, the U-tubes are rapidly cooled ater a leakage rate has In this way, the meroannulus is bel ‘sealed by the shrinkage ofthe U-tube, and the ‘subsequent gas flow rates canbe wholly ateit- ted to matrix gas channeling. This apparatus measures the pressure dilleren- tial arose the U-tube. In tis way, ican monitor ‘the prossure-transterrng capacity ofa slurry due- ing gelation and during setting. the pressure ‘ifereniataeross set cements es an the pres- ute imposed originally, the cement has leaked, and Shol ean then calculate its permeability. ‘Tho experiments start with the same pressure ‘on both sides of he U-tube bocause the cement is liquid ad able to transit pressure, But, once it starts to set, a illerential is established the ‘slurry sable to malta hs diterentil, Shell believes thatthe cement will be ale to stop, or at east delay, gas migration. The selection of apparatus to testa slury effectively slects simplifications which in ura help determine the ehoie of additives and job procedures. The complete simulation of a well, with gas intrusion capability remain a dream Utube gas sue- ton apparatus for Presewe measuring cement transducer Permeability 10 ‘gas. Shell mea- Sures permeability ‘by applying a ait {orential pressure ‘across the tube. 250m Utube inqven at 20-100" 39Commentary: Hal Grant Hal Grant, supanisor, cementing servoas sect, Chevon Series 60, flow potential here.’ If we're involved in designing the slurry, we'll allow for that possibility from the start. Everytime we cement a production string Hiner, nd anytime we set pipe across a possible 428 bearing formation, or even a saltwater forme tion, we take precautions against gas inf. Wo try to keep our approach simple. We tink there is only one thing that wil et ow happen, and thats loss of hydrostatic pressure, There are eran propertos that we like to put into ou slur that pres if to developing initial strength. Good ‘ensiy control should ensure that you're actually mixing the designed density. In adlton,well= ‘known good cementing practices, lke centraiza- tion and movement of pipe during a coment jb, ‘should be employed. ‘Of these fe, four can be Incorporated int the ‘design by doing some laboratory testing and using selected adtves. The ith, density conto, is ‘something that has tobe done on location, ding ‘the mixing on the job, andes in the lp of he service company. ‘We nearly always recommend these properties, but sometimes the service company comes up wil recommendations. We listo and may use some of ts adatves to enhance these properties. 40 Putting Theory into Practice Current solutions to stop gas intrusion fall Into three categories: optimization of con ventional good cementing practices, adding ‘new properties to cement slursies and hal \ware/procedural mocifications. (Of all cementing practices, effective mud removal is one of the most important and hhas a major bearing on preventing gas int sion, Field experience shows that however {ood a particular slurry is at resisting gas Invasion, reliable zonal isolation is possible only when coupled with effective mud removal practices. Prevention of mud stringers depends on completely displacing drilling fluid during, cement placement (see “Mud Removal: Research Improves Trad: tional Cementing Guidelines,” page 44). ‘Another aim in conventional slurry design, particularly for cementations requir- ing maximum integrity (like lines), is con- tuol of fluid loss from cement into the forma- tion, Since reduction in cement pore pressure is proportional to the reduction in slurry volume, fluid loss control is even _more important when gas intrusion is antic pated. In the laboratory, increasing the con~ centration ofthe lui loss control agent and ensuring a low fluid loss test has been shown to lead to less gas invasion and lower cement permeability In high-angle wells, microannulus devel ‘opment has been attributed to settling ofthe slurry and release of free water. This is because, in some circumstances, the free ‘water can form a continuous communica: tion between a gas zone and the surface (above, righ). Slury stability, a goal for all ‘cement designs, becomes even more impor- tant when gas intrusion threatens.” However, such improved practices alone ‘are often inadequate to fight gas intrusion, Isolation can then be achieved only by spe- ial modifications to the cement slumry (see / Hal Grant” (eft). Special ys- tems reflect the diversity of approach within the industry. Gas bubbles, rapid setting and ‘modified gel strength development have all been used. More recently, impermeable cements, those containing surfactants and slurries with viscostied mix fluid, have been added to the list of options. These additives all aim to prevent channeling through the slury matrix. Expansive cements have been sed to combat microannulus development. [)Mhe importance of stability a slurry used in @ deviated well is unstablo,tr00 water can form a channel on the high Side of the wellbore. Enrained Gas 'Notmal cement slurries have low compress: ilies, so small volume losses create sis ificant pressure reduction in the annulus. Slurries containing entrained gas are designed to be more compressible. As ‘cement hydrates and shrinks, enlrained gas bubbles expand to compensate for the vol- lume reduction and so maintain a positive pote pressure across the formation. These slurries are produced either on surface as foamed cement by the introduction of an Inert gas (usually nitrogen) or in sit through use of additives (usually fine aluminum oF ‘magnesium powder) that react with the cement downhole to produce hydrogen, Which creates a sort of foam, Deliberate introduction of a gas into the slury seems to be a perverse way of oppos- ‘ng unwelcome gas entry. However, the gas {in foam is not mobile because it forms small bubbles that do not coalesce. But foamed ‘cement’ low density and high compressbil ity can lead to difficulties achieving suffi lent hydrostatic pressure to place the ‘coment in deep high-pressure wells without collapsing the foam, Some success has been reported with the insitu development of hydrogen Because hydrogen is explosive, questions have been raised over the safety of operations using hhydrogen-producing additives, both during slurry placement and when the set cement is dilled out. Further dlfculies can arise the generated gas breaks out and migrates upwards, enhancing the chance of gas nflow rather than its restraint? Oilfield ReviewRight-Anglo Sotting Rightangle setting cements, named for the shape of their consistometer readings in the laboratory, show no progressive gelation and set very rapidly (below). These proper ties allow transmission ofthe full hydkostatic load of the cement colurmn to gas-bearing zones until shortly before the cement sets Reducing the slurry’ transition period cuts the time available for gas intrusion.! Delaying development of gel strength has the advantage that hydrostatic pressure is maintained while any reduction in slunry volume due to fluid loss is occuring—an tungelled column can more easly compen- sate forthe volume loss However, right-angle setting sluries tend to exhibit high volumetric shrinkage. Fu theemoe, if the bottomhole circulating tem: perature is less than 121°C [250°F), itis hard to design right-angle set slurries because hydration kinetics of cement are significantly slower, Below 27°C (80°F) plaster can be added to cement to give right angle set properties. 00 Comoressive strenath, psi & ° fe ro patod goo 8 Tine, rin Rightangle setting coment. By moving zapialy from liquid fo sold, rightangie setting cement spends the minimum time ‘possible in the vulnerable transition state, Jn this case loss than 10 mlnutes. Consistency and compressive strength nt fests wore carried out at 16°C [60"F] and atmospheric pressure. “Son DL, Sabi Fad Fal “Annular Gas ow Ther and Prewetion Maths Dever Os ‘nd Gs foual 2 (Decener 10, 1984 8492 6 Stewart RB an Schouten FC “Casson an ‘Migaton in Caren Arl Causes a Ure” paper SPE 1477, peor 1986 1ADCISPE Dring Coherence, Oa Tous, USA, Februny 1012, 8 Sour DL and Wilkinon J: "Combating Gas Mi tin nthe Mihsan t” pape Pe 19314 ye ‘ented atthe SFE Eston Reson Meet, Maan ‘on Wes Vin, USA, October 29-27 1908 Cisin WW, Chater and Osrot GA "Cas Leakage in Pinay Come AF Stay ad Libostn lestation four of Polen Techno 28 (19782 13611508 Apvil 1991, Problems that arise with cement jobs co not necessarily need new t. ipplying current technology correctly. The challenge is ogy. We may not hol to make people in the field understand what is important and appreciate the need for quality control during cement job: ‘Thore isa whole list of points to which you must adhere and if you do so am confident itis possi- ble to suszeedin 90 percent or more cases. This ‘requires an understanding ofthe mechanisms at play, cementing program that i“ for purpose’ ‘and good quaiy cont For many years we have stressed an integrated approach. Using special aditves is fine, but it your displacement is poor, you are not going got anyuora. We belive tha this gnera ‘means using atin scavenger slurry ahead of 2 ‘main slurry wth fou td os, displaced at as high 2 pump rate as possible as zone cementallon equies the same rules ‘ot employs extra aditives. Thee is potential for gas to migrate through the eoment —In add- tion traveling through a microannulus ora channe!—and we realize that in certain cases we ara require to stop it ‘nthe microscale, gelled cement can be con- sidered 2 matrix into which gas cane experiments, cement placed inside a steel pipe ‘ated internally wit ubber to avoid erzation of a ‘miroannulus and exposed to diferent pressure has detntely exhibited flow though the m not arg volumes, but ina well enough to ereate ‘significa pressure atthe top ofthe annulus. ‘Supporting this isthe observation tha al the ‘successtl gas migration adtves—atex ‘cements, adatves with a foaming capacity or Imerastea—plug gelled coment’ relatively caarse porous matrix. This suggests nt only that ‘the mati song enough to hold the plugging ‘agents but that gos migration occurs though the mai and not just by macrosized blobs of gas. Higher volumes of yas ow indicate mad chan nes or microanul. To prevent the formation of 2 Imeroannulus ater the cement has set, we 1 Webses WW and kr Flow Aer Comet ing A Fell Laboratory Sty” paper SPE 0259, fresnel at he a SPC Ara Teh Cone ‘ce an shin Las Uae, New USA, Sipe 3316, 1999. 9 awe DR Hato ad Ste: “A neat ‘proach fr suetenfl Primary Cement” piper P2959, presen Ste tO Techn x Conerence; Rana, Bain, Mach 12, 101 Rang CLTvauaton of Gas Flows in Cement bbe SPE 16385, peso Cai Reyna eee, Vera Calf, USA, Api 10. 1987. Gerard Bol, section ade, deg ‘ids and coment, Korine Exporate on Produkt Laboratoium {woke closely at seat rings placed on the outside ofthe casing Full evaluation in the felis it- cll cause we can never be sure whether the seal ring has performed ot whether there was no ‘mieroannuls anyway, In aboraory experiments using sea rings, we found that there isan enor- ‘mous improvement ‘The mechanisms clear: if you havea casing lundr pressure and if at any time during the ite of ‘he wall the pressure is et ot, 2 meroannulus ‘may develop. So any well that you expect to art: clay itor one that wil show clear reservoir pressure drops may need seal rings. However, withthe exception of the Gemoco seals, there are no purpose-esigned seals avail eHow. We stared to de yetto come to futon. We intend to review tis to see if any further ation is required, ‘Modified Gel Strength Diameirically opposing the above philoso- phy are thixotropic cements that rapidly develop gel strength which can be broken down by shearing, In this case, the gel strength is designed to prevent the flow of {gas bubbles before the cement sets. This gelation can be achieved through creation fof a microcrystalline structure or mictoge: latinous network of mineral hydrates throughout the coment slury. Alternatively, polymers dissolved in the interstitial water can be chemically treated to yield a sel supporting viscous gel Such systems are reported effective in combating entry of gas bubbles.» However, it is debatable whether they resist gas ‘molecules that are smaller than the cement pore size and can enter the cement matrix ‘once it has become a twophase system (see “Commentary: Gerard Bol,” above). 4Pormeabity Blocking ‘As cement makes the transition from liquid to solid, it becomes a tworphase permeable system. One way of preventing gas migra tion, therefore, is to reduce matrix perme: ability.!! Several materials used for this include microsilica and lates. Microsilica comprises 85 to 98 percent amorphous silicon dioxide (SiO,) in the {form of small spherical particles 0.02 to 0.50 microns (ym) in diameter—around 100 times finer than cement particles. The liny particles are thought to plug spaces between cement particles and reduce matrix permeabiliy."2 Another explanation holds that microsiica’s very high surface area—15 to 25 square meters per gram (m/e) com pared with about 0.3 m4/g for cement ‘means that it has high pozzolanic activity and reacts with the caleium hydroxide in the cement, futher reducing premeabilty. evelopment of the microslica additive centered on cement systems lighter than 14 pounds per gallon (lbm/gal to control shal Tow gas pockets encountered in the Gulliaks field in the Norwegian North Sea. Now den- sities range from 11 10 16.2 Ibmigal. The low densities also use glass beads as a lightweight extender. Large volumes of the additive are rneeded—more than 20 percent by weight of ‘cement is often recommended. This can present bulk handling headaches and has also raised doubts about how such large ‘volumes of adcltive affect cement strength. This is particularly true at high temperatures at which silica fume has to replace the larger silica flour which is traditionally used to oppose steength retrogression."" Using both additives would leave too litle cement inthe slur. Another permeability blocker is latex. Solid latex polymer particles are dispersed in water and stabilized by anionic surfac tants and protective colloids. The polymer particles are diserete and remain dispersed wathin the slurry until setting stats, They are small enough (0.2 to 0.5 jum) to occupy the cement’ pore spaces, During setting, the pore water reacts with the cement and the Tatex paticles coalesce to form a coherent, low-permeabilty plastic film that blocks fur- ther gas migration through the cement ‘Although many latexes are commercially available, only a few can withstand the drastic requirements of storing, mixing and pumping, and then the downhole temper ture, pressure and chemical environment which can cause flocculation and loss of film-forming ability. Because ofthese sensi- tivities, stabilizers are also employed.'¥ 4a Surfactants and Viscositiers ‘A technique in between entrained gas and permeability blocking uses the foaing capacity of some cement additives. Some standard adcltives like dispersant offer lim. ited foaming, but for greater foaming capa- bility, surfactants are added to the cement slurry. When gas enters the matix, the sur factanis react, forming a stable foam at downhole concltions, These stabilized bub bles can then oppose matrix influx in the same way as foamed cement. This mecha nism has been confirmed in studies by Koninklijke/Shell Exploratie en Produktie Laboratorium, Rijswijk, The Netherlands \Viscosification of the interstitial water is yet another technique, This seoks to create a gummy, viscoelastic film within the cement pores (below, limiting the mobility of the ‘cement filtrate and reducing gas mobility (see “Commentary: Franco Marcassa,” next page).!s Viscosifying polymers have also been used along with a blend of coment and slag, {containing primarily oxides of calcium, sili- ‘con, aluminum and magnesium) which, among other attributes, helps reduce slurry porosity further impeding gas intrusion. 16 ‘Another idea related to viscosification involves reducing cement matrix pore throats using bridging and polymeric agents, which swell to further immobilize fluids in the cement pore spaces. Cattonle polymer in action. Within setting coment cop), he pore space foxcoeds 2 microns in diameter. This ‘space has been filled dbottom) by {dalton of 15% by weight of water of a ‘cationic polymer that increases in ‘Concentration as cement hydrates. By {ling the spaces, the polymer prevents ‘gas migration through the matrix. ive shrinkage, % E 6 Time, br \Coments shrinkage with time. After 60 ‘hours at 65°C (150°F] and 3,500 ps, this Class B neat coment has shrunk by neatly 4%. Expansive Cements Expansive cements aim to prevent the for mation of microannuli by opposing shrink= age with cement expansion. Cement sys tems that incorporate gas bubbles mentioned above are olen quoted as offer ing expansivity. However, once the cement has set, entrained gas can no longer expand And so does not offer much benefit (above). The other principal method for achieving ‘expansion is through crystal growth. Etrin= site—a highly hydrated form of calcium sul- foaluminate—may be formed by the reac- tion of calcium sulfate with the aluminate phases, Such reactions give a linear expan sion of 0.2 100.5 percent but are limited to ottomhale static temperatures of 85°C 1185°F}, Expansivity can also be attained through addition of high quantities of sodium chloride, sodium sulfate or both, ‘Addition of calcined magnesium oxide at a concentration of 0.25 to 1 percent by ‘weight of cement also provides an expan: sive force within the cement due to hydra tion to magnesium hydroxide. Expansion increases with increasing temperature but proves insufficient at less than 60°C 1140°F].!7 However, it can lead to low strength in set cement and, when greater than T percent, expansion can leave cracks and fractures, Oil companies are research: ing ways of combining gas-blocking proper ties with expansivity. Oilfield ReviewHardware amd Procedures ‘While slury modification techniques have become an important element in the fight against gas inttusion, earlier efforts concen- trated on use of special hardware and modi {ying operational procedures. '®19 Physical means to reduce gas migration include using external casing, packers and stage cementing to cut the height of the cement column and effectively reduce the amount (of hydrostatic pressure that can be lost dur ing setting, In the past, much emphasis was given to increasing mud density in the well prior t ‘cementing; today, this is less popular, though ome operators recommend pressuring Up ‘on the annulus during cementing. Mechani- cal seal rings have also been employed to help seal the casing-cement interface and ‘guard against the consequences of casing contraction. Furthermore, seratchers mounted on the outside of the casing have been used successfully to remove mudcake ‘on the formation and improve mud removal ‘Annovel way of preventing the buildup of gel strength and therefore avoiding the downhole pressure drop involves vibration fof the casing.2? This has been tested in both the laboratory and in a test well. Favorable results are teported with the additional ben- efit of seemingly improving the hydraulic seal atthe cement-formation interface. —CF 11 Sep Oa Cat 8 “New Quick Stig Cement ‘Soh shallow Cas ation bles nd duces WOC Tine” per SPE 1500, reset lhe SP Este Reool Metin. Manolo, ‘Wes Vipin, USA, Novena 8,198. ‘yes RL and Log New Tchoogy in Gas Maton Corto gaps SPE 16653, peso a thecard Pt arma echnical Conerence ad Fshitlon. Dale, Teas, USA september 2740, 12. Banner CE, Shute GF, Wooded LA, Jones Md Sn Bh AG "Crt! Deagn Parameter ‘on as asian During Cementing Operations” ioe SPE 11982, peered a te SB SPE Annual {echncl Conerence snl Eaton, San Franch, Gita, USA, October 58, 1983, 13. Gro assy B and Digs £0, “Dewey ‘nora and Uw of Gag Comet” paper {ADCSPE 17258, presented at he ADCS Dain Content, Del, Texas, USA Feary sarc 2 1868. April 1991 Commentary: Franco Marcassa For us, gas channeling is an old story. In 1976-77, offshore Borneo, Indonesi ‘wo oxporionced gas at surface and lost ono two wells. At that timo, there were no commercially available additives to combat gas intrusion, so ‘we in research were asked to find some way to control the gas. One of the main driving forces of gas chan Ing isthe pressure oss that comes from the sll. ‘suppor ofthe coment column which reduces the hydrostatic pressure exerted onthe formation. ‘But don't forge that seting cement contains a network of small hoes. This edt the develop- ‘ment of a cationic polymer that increases the mix ‘id viscosity asthe cement sts, eating a vis- ‘ous material inside the porous matrix ofthe ‘ament fo oppose eas migration, This resstanco oes nat depend on pH or any ater additive, only ‘on the water conceatalon, which is cotalld by setting, There is an optimum polymer concentra- ton to make a cement gas-proot. This eatonic polymer has been used on more than 100 wells ‘ith 95 to 98 percent sucess. The key othe viscosity increase isthe loss of ‘wale ino the formation. This loss aecus during hydration and increases the relative polymer con- ‘contration and hence the tues viscosity The mix ‘alr viseosity of about 30 contipoise (en) wil vise to as much as 5,000 ep during setting. One aspect o the polymer shat it tends to make the ‘slurry viscous This ean tim sary flow rate dut- ent in cartain cement obs. when you ad a produett st ‘025, people tend to think that it wil stop gas whatever you do, Ths is wrong. Al products for stopping gas migration are designed todo so Inside the coment. Gas channeling may be ‘sult oan operational problem arising from bad ‘casing cleaning, a permeable zone or bad opera tion sequence, 14. Grabowsh ad Gil “et of Replace ‘Sica our Wi ies Fume on Eoin, Froperer ot Ofhell Cement at Noma a Te ‘ied Temperatures and Pesste” Cement nd Conerte Reser 12199153348 elon EB ed Comet. Area, The Nether Ever Science bse 1950. 15. Drag P and Parca PAA Single Techie Sah Cae Migration Peblems Ass Wie Rare o! Cnction” paper SPE 7029, resented Se SPC cata necting ethers ‘cig Hani, Ching, November Io, 188 Chg PR aa te RM “Ca la in Ces” fae SPE 11207 peso at heh SPE Ans TechnealConrence a ation, Net Oca, Lubians USA, septobe 1629, 1982 Cheung and Mick HO Field Gaston ofan Inparmeae Coren ten fr Corin Gas tation paper SPE 1183, present he 3h ‘nial curl Canerence sa Exit, Sa Francisco, Callow, USA, Doeber 5 183. 17. es at Rone improved Gas Maton Con telina New Ou Wl Cement pape SP 17926 [resent tthe Mul fat Cecil Cane: ‘ce and Exhibit, Nona bain, March vie, Franco Marcasse, senior dling fid and coment engineer, Total Compagnie Francais es Patoes, Conte ‘Scieniqu Technique cause we are dealing with alot of money In {he eld, people want to save time, and so when e cement job i tnished they want to str nip- ling up and testing the blowout preventers and pressure testing the casing If catried aut oa soon, ‘hese operations may induce gas channeling. And you have to et the ob right he fst time. 1 oot believe too much in remedial cementing, especially when there is gas channeling. The space {tobe squeezed i ton smal even for extr-fin ements, This fs because you donot know where ‘the squeere cement wil go. When you have chan- noing problems, you pertorate the casing and gt erally squeeze most ofthe cement into the forma ‘lon, not the channel that needs i. The chances of icant quantities of cement going into the el ae veal al ‘ith gas, as soon as you havea small leak itis forever. Is not ke liquid. You can see this nthe 92s equipment 1 Seidl FA are Gece 1G: "Use of Esprsg ‘Cement nprowes hing ard Ald eating, ‘nul Meson ind qb Sa ‘Anes Wel paper 2 1441, preset a he (a Pt Arua etal Conese a si ton, Lass Neva USA Seem 2225, 1905, 19. esne DG, Thos, ezner HP and Tolle Gc “fanaa Gas ow Ader Crest A Look at Pa ‘al Solus” paper SP 8255, pce tie Seth SPr Arma eens Coren a Tah thn, as Vigan, Nevads, USA, September 13.25, 1990, 20. Cooke CE, Gonzales OF a Brows OF “Pinay ‘Cement mprovemen by Cig WEration Da ‘Sed ths NSPE wu cial Coen Bas ose 4B