RESERVOIR ENGINEERING Henri Freyss Houston Paul Guieze Nikos Varosss. ‘Melun, France ‘A. Khakoo Karen Leselle Dick Simper ‘New Iberia, Louisiana PVT Analysis for Oil Reservoirs ‘The challenge for reservoir and production engineers is to maximize hydrocarbon tecovery in the minimum amount of time and with the smallest expense. In terms of the engineers’ on-the-job concerns, this ‘means asking a variety of questions about what and how the reservoir will produce: How lange are the reserves? What will the primary recovery be? What kind of crude will be produced and what will the market pay for itt Does the crude contain unwanted com- pounds that will destroy tubing or plug the weil? For offshore wells, will the crude solidify at ‘acean-bottom temperatures, thereby stop- ping production? What separator pressures will maximize lia- Lid hydrocarbon recovery? Will gas produced at the surface be reinject: ‘ed, sold, or burned off? Answering these questions requires antici pating the volumetric and phase behavior of produced hydrocarbons as they travel from the reservoir, up the tubing, through surface separators, and finally into pipelines; and of hydrocarbons in place as the reservoir pres- sure declines with production." Hydrocar bon composition, viscosity, density, com- pressibility—all these properties change with pressure and temperature (see “Intro duction to Hydrocarbon Phase Behavior,” age 6). In a pressure-volume-temperature (PVT) lab, researchers employ an arsenal of instru- ments to determine reservoir fluid behavior and properties from oll and gas samples. Their goal i to simulate what takes place in the reservoir and at the surface during pro- duction, A central aspect of PVT analysis is High-pressure, high-temperature pressure-volume-temperature (PVT) cel The smaller cel Grelined eylinde i used or for bubblepoint determination, liferental Horatio tests, 2nd Mash liberation tests on ol reservolr samples. tis rocked back and ferth before each fash and diferent liberation. The Tang cell vertical eyinde is used to recombine srtace ol and as a. ples at reservoir conditions how gas evolves from oil when pressure falls below the bubblepoint. Researchers use two processes to describe this—flash liberation and differential liberation (see page 9, middle In flash liberation, a sudden pressure drop causes gas to come out of solution, which then remains confined with the oil that Is. left In differential liberation, gas comes out of solution as pressure is gradually decreased, but is removed from the oil ‘When reservoir pressure drops below the bbubblepoint, flash liberation dominates, Gas ‘comes out of solution and intially remains The Technical Reviewinthe pores in contact with the oil. Mean- while, single-phase oil continues to flow Into the wel, although it has a slightly dif. {erent composition than that ofthe original reservoir fluid. When enough gas accumu- lates to reach the critical gas saturation, it begins to flow, Since gas is less viscous and flows faster than oil, the gas liberation now becomes diferental. Gas liberation occurring in the tubing is considered flash liberation because the agi tation of the flow keeps the two phases in Contact with each other. Flash liberation aso dominates in the separator, where a sudden pressure drop releases more gas from the oi while the two remain in contact, Flash and differential liberation are recre ated on a small scale in the lab using high: pressure, high-temperature PVT cells, posi tive displacement mercury pumps and precision thermometers and pressure gauges (left. When introduced to a sample-filled cell, mercury compresses the fluid; when withdrawn from the cel, it allows the sam- ple to expand, ‘The flash liberation testis used to deter- ‘mine the bubblepoint of the reservoir fluid. Both flash and differential liberation tests establish parameters that relate the surface volumes of produced oil and gas to their reservoir volumes. Bubblepoint Determination The bubblepoint is determined ducing a flash liberation test on a small amount of reservoir fluid (about 100 cm’), sampled downhole or recombined from surface gas and oil samples taken from the separator. ‘ter transferring the sarmple to a high-pres- sure cell maintained at reservoir tempera~ ture and above reservoir pressure, the cell pressure is decrementally decreased by removing mercury from the cell. For each volume change, cell pressure is recorded Volume 37 Number 1 Upon reaching the bubblepoiat, gas comes ‘out of solution and any subsequent mercury removal causes small pressure reductions. This transition is well marked on a pressure ‘volume graph of the data (below. The bubblepoint is crucial for understand ing how hydrocarbons behave in the reser ‘voir and indicates the probable drive mecha- nisms. For example, ifthe reservoir pressure is initially at the bubblepoint, a gas cap almost certainly exists that wll exand, push ing oil toward the well and increasing recov- ery t0 40 percent of the ol in place. Ina reservoir producing below the bubble ppoint, gas coming out of solution may either be produced, or if production is slow and vertical permeability i high, it may migrate ‘upward and form a secondary gas cap that helps deve the reservoir. This type of drive, Pressure —= Bubble Point A Volume —> [iBubblepoin determination ding a Ash ee ation fst. Staring above the reservoir pressure, ressure onthe reservoir sample i dereased io Steps by romoving mercury fom the cll. The n= {ial slope ofthe pressure-volume cure is very Stoop uni gas bugns fo come out of solution at the bubblepont. The, any subsequent volume Increases cause small changes in cel pressure and the lone ofthe pressure-volume respons Shows a dramatically reduced slope called gravity sogregation, may yield up to 70 percent ofthe ol in place. initial eservoir pressure is much higher than the bubblepoint, meaning the oll is undersaturated (contains litte dissolved a8), then the only dive will be monophasic liquid expansion, resulting in poor recov- ‘ery—only 2 few percent ofthe ol in place. ‘An enhanced recovery program may then be implemented. The presence of a natura ‘water drive in this case, however, could lead to a recovery of up t0 6 percent of the oil in place. In the case of a retrograde condensate reservoir, the only drive mechanism is gas ‘expansion, but as pressure declines farge ‘amounts of ol, called condensate, may con. ‘dense from the gas and remain in the reser 1 Recommended reading and viewing 0 eer erence ‘mp f Bas DSU and Whiting RL: Pete ‘sero Eginrng. NewYork MEG Book Conan, 1360 (Oak Us Fusment of Reser Engng ‘Amsterdam Esco Scene Potihng Company io Lar KR: sent Fads: amiga Aral ‘Mode Pee: Peoleun Eien IRD Video tay expan and Podcton Specs ou neratonl Haman Resources Deveopvent Corporation 1985Introduction to Hydrocarbon Phase Behavior To understand the complex behavior of a reservoir fluid, le’ fist follow the case of a single-compound hydrocarbon initially in id form in a pressurized cell containing mercury (below). As mercury is gradually removed with the temperature held con- stant, the liquid expands and the pressure in the cell drops. Isothermal conditions simu late areservoir’s generally constant tempera- ture. The pressure continues to drop with additional mercury removal until the first bubble of gas comes out of solution—the bubblepoint. Below the bubblepoint pres- sure, the gas continues expanding, as mer- ccury is removed, but the pressure remains Constant. When all the liquid has vapor ized—the dewpoint—further gas expansion is again accompanied by decreasing pres- This experiment run at different tempera- tures produces different bubblepoints and Hn preesre Renter aeetneenceane eater tree ripe eee eet bor showing be enpoarand presi condtrs at prod ga Ligua Gr ladda Ate pesare sd enpcaure Coneponig he ot Hescarnon Supine gr and hd hae a deigsae Al Liquid Ree Phase behaver ofa dewpoints. Projecting the locus of bubble- ‘single-compound points, the bubblepoint curve, and the locus ‘nydrocarbon in a con- of dewpoints, the dewpoint curve, onto a Ae sar tempat essureteperate graph Yields a ures Ft Ga erst reels viding line known as the vapor pressure hydrocarbon expands curve (above). Above this curve, the hydro- . andthe cellpesie carton exists 8 ul, bas 9s. At dons This continues the critical point, gas and liquid phases are tothe bubblepoiny, aes iaserdlnid Gas when the fst gas bub in biecomes outof solu. Multicompound hydrocarbons generate tion Below the bub: complicated phase diagrams, whose shape G bepoint the volume depends on the composition of the mixture Bakes increase @pil¥’ right, above). The bubblepoint and dew- butt consan pres Sureunifaof¥e ge point curves projections no longer coincide, Last Droplet, tit has vaporized, a" but produce an envelope in which gas and tedewpoin Below liquid coexist The maximum temperature of F ESSIEN sSevciacencipesaleltecio creasing presure. Phase diagrams illustrate the evolution of hydrocarbon phates during production Pressre depletion represented a a en Calne because he ferovol temperature E ‘generally remains constant (far right, top). ‘All Gas The Technical ReviewPressure: Temperature ———e Phase aiggram ofa muticomound hycrocarbon mixture Within the ‘envelope formed bythe bubblepoint and dewpoint curves, the fluid exists as _g38 and Tiqut, above the bubblepoit curve as only liquid, ane blow the Pressure——+ Pressure—= Retr ‘Condensate Reservoir Temperature»Roleased and Greatly Expanded Gases oi Oi Produced o Surface Undergoes Pressure and “Temperature { Reduction Released ‘Separator [LiPresure changes encountered by reser uid during production that result inthe iseration and expansion of gas and the shrinkage of ol. Aer Lang KR, reference 1. Pressure, KPa (x 10°) 0 o 0 0 0 9 = 2x0 fa 380} % = 1,200] 28 xo] 8 2 § ua 2 2504 B 3 0 za 21 § = a) 20S $ = m3 5 & & 1] § 10 a o wo] & all he ° aaa aaa Bane BIT Prsoue, pia The il eration volume actor By ae solution gal aio Ry 28 fnc= tions of pressure. Theol formation volume factor account forthe fact that «il below the bubblepovn iberates gas downhole resulting in les at the Surface. Always greater than one 1s the ato of the volume of a quantity of ‘reservoir oil vied by the volume of the same quanity of ol t standard ‘cnchtons, The solution gas-ol ratio tells how mach gas issued nthe reser of Reservoir-to-Surface Volume Relations Besides the bubblepoint, there are three important parameters derive from flash and dllferential liberation tests that relate surface volumes to reservoir volumes and thus help determine the amount of hydrocarbons in place (ef. These are the ol and gas forma~ tion volume factors, B, and By, and the solution gas-il ratio, * B,—The oil formation volume factor {equals the volume of oil that must be with- drawn from the reservoir to produce one barrel of stockctank oi? at standard surface ‘conditions of 14.7 pounds per square inch absolute (psia) [1 almosphere] and 60°F 115.5°C]. Itis expressed as reservoir barcels per stock-tank barrel [tes bbU/STB or res imAtocketank The olin place in SB is V9 (15) crores where geologic, logging and core analyses provide estimates of reservoir volume V, water satutation Sy, and porosity ¢. This manner of estimating reserves is called the volumetric method. + By—The gas formation volume factor ‘equals the volume of gas in the reservoir that will produce one cubic foot at surface tunder standard conditions. tis expressed as reservoir barrels per standard cubic foot (res bbbVscf or res m¥std m'). The standard vol- ‘ume of gas in place is: VOUS) A + R.—The solution gas-oil ratio isthe vol lume of gas at standard conditions that di solves into one STB of oil under reservoir conditions expressed as standard cubic feet per stock-tank barrel [sc¥/STE or smi/stock- tank mn). The amount of gas ins i Vo USK Bo : Once production data become available, the formation volume factors and the solu. tion gas-oil ratio may be used in more sophisticated estimations ofthe hydrocar- bons in place using material balance equa- tions. These account for water entry, the expansion ofreservoit lids and the redue- tion in pore space, all consequences of falling reservoir pressure. ‘The formation volume factors and the solution gas-il ratio, all functions of reser ‘voir pressure, are determined by reproduc- ing in the laboratory the sequence of difer- ential and flash liberation that most likely ‘occur during production (lef) Itis widely The Technical Reviewbelieved that flow in the reservoir is best simulated by differential liberation while flow up the well and through the separator is best simulated by a series of lash libera- tions called a flash separation, The laboratory set up for a differential lib- eration testis identical to the flash liberation test, excopt thatthe pressure reduction stats from the bubblepoint, and the gas liberated after each pressure decrement is removed through a port while mercury is injected at ‘constant pressure (right). At each stage, the {ree gas and reduced oil volumes are mea- sured as well as the free gas volume at stan- dard conditions. Eventually, after four to ‘eight pressure reductions down to atmo- spheric conditions, the cell contains only residual il at reservoir temperature, which is then cooled to standard temperature and has its volume measured. ‘A lash separation test is used to establish ‘optimal pressure setings forthe surface sep- arator and the appropriate number of sepa- ration stages, both of which are designed to keep the light hydrocarbons in the liquid phase and maximize liquid recovery. The test involves a series of flash liberations of a fluid sample injected into a small-scale sep- aration system. The separation pressures and temperatures are altered in steps, and at ‘each stage, the volume of gas liberated and the volume of liquid remaining are mea- sured. Data from these tests are combined to determine B, , By, and R, that reflect ‘actual production. ‘Compositional Analysis ‘Compositional analysis ofthe reservoir sam- ple, a key component of PVT analysis, has several applications in reservoir and pro: duction engineering. The most important i to establish how much gasoline, kerosene, fuel oil, heavy oil and bitumen will come from refining a barrel of crude, which deter- ‘mines the crude’s market value (see “Com- 2 Scns ol or conden collected na Tae storage a a aimzspr pressure ae ne produced wel fdas been processed ough a {Gecal separation ss. 3 Fora resem of gs cvomaagaly: ‘Menai and Sonal sie Cas Onomatoge- ‘hy Palo Ao, Caller Varn ister! Dion Biices 1969 Volume 37 Number 1 Flash Liberation Below Peuttiepoint —Pbubtlepaint Differential Liberation Poadbecins Bolow Prsbeiepoint [Comparison benwcen fh iteration and iter ‘ental Hberaton. Fash hiberation, which portrays {he separation stage of production, voles» sudden pesure change tha cases efi to ‘chy Separate o% “sh I asad of com ponents in iferental bration, which ope. Sens peste depletion nthe reservoirs ‘raaly decreases and any liberated as femoved fam eo position of Hydrocarbons,” page 12). The ‘composition also dictates how and where the crude will be refined. A light crude can be refined through distillation, but heavier crudes must be processed at very high tem- peratures (1,000°F (540°C), which break or crack lager molecules ofthe paraffinic series into smaller, lighter molecules. Refineries configured to process a specific composition of crude economically may not accept al types of crudes. Another applica- tion is detecting corrosive compounds that require special noncorrsive producing and ‘ranspontation equipment The main compositional analysis tech- ria is gas chromatography, In ths tech- nique, a hydrocarbon sample is heated to ‘vaporization and circulated by an inet car- rier gas through colurmns of porous adsorb- ing material (next page, above. The adsorb- ing material may be a solid—silica gel, charcoal or a molecular sieve of synthetic zeolites —or more commonly a liquid sol vent that coats particles ofan inert solid. The sample is introduced into the injec- ion port of the chromatograph using @ syringe or a gas-sampling valve. The port is at a high temperature—650°F to 850°F [350°C to 450°C|—to vaporize the sample before it is injected into the columns. An inert carrier gas such as helium, nitrogen or hydrogen circulates the sample through the adsorbing columns. The individual hydro- ‘carbon compounds alternate between being, liquid, solid or gas. Asa liquid or solid; they remain adsorbed, while as a gas, they are carried by the inert gas out of the colurmns. Lighter compounds preferentially remain gaseous and are quickly removed, whereas heavier compounds progress slowly through the column. Fach compound is identified by ‘observing how long it takes to ex The compounds are usually identified by a flame ionization detector (FID) or a ther- ‘mal conductivity detector (TCD). Inthe FID, a flame composed of hydrogen and alr burns off the exiting compounds, fonizing them. Changes in the flame’s conductivity, ‘proportional tothe ions and electrons in the ‘compound burned, show up as chromato sraphic peaks. The FID detects compounds heavier than C,H,, or in PVT shorthand Coy (see “Composition of Hydrocarbons," page 12), Unfortunately, the FID technique destroys the sample, and cannot recognize ‘compounds that don’t burn, like nitrogen and carbon dioxide. ‘The TCD detector measures the character- istic ate at which heat dissipates through each compound. The heat from a filament of platinum, tungsten or nickel with constant ‘current running through it is dissipated at a ‘constant rate by the cartier gas. The rate of heat los is reduced when a mixture of sam- ple and carrier gas passes over the filament, ‘causing an increase in the temperature and the electrical resistance of the filament. The ‘changes in resistivity are recorded as peaks. ‘The TCD detector can be used for analyzi ‘compounds up to Cs oF Ce,Sample: JColumn Cross Section Injection Pst Detector Gas Chromatograph Chromatogram lament Thermal Conductivity Detector (TCD) Gas cromtograph fr compositional analysis of resrvor ids, reservoir samples inacted and ‘vaporized at high temperature, then circulated by an inet gas such as helium through columns filled {with adsorbing mater. Inside the columns, te compounds inthe sample alter Detwecn the car: lerphase and the phase ofthe ackorbing materia (either solid or liquid. Eventually the compounds Ink ther way ot ofthe columns to be detected, with the lighter compounds eluting fst. The com pourcs ae Getcted by either 3 Tame ionization detector (1D) oF a thermal conductivity detector {1CD). The resulting output of peaks splayed versus tne is called a chromatogra 10 ‘The resulting output of peaks from either detector is called a chromatogram (above right. The area under each peak is propor tional to the concentration of the corre- sponding compound. The time it takes each ‘compound to pass through the columns is knowin as the retention time. Retention times ‘and peak amplitudes are calibrated by injecting the chromatograph with a standard mixture of known composition, ‘A limitation of gas chromatography is that it cannot make a complete analysis of stock-tank oils that contain heavy alkanes and asphaltenes, which take unreasonable amounts of time to elute. To analyze these heavy liquids, you can inject the stock-tank oil into the chromatograph while keeping ‘out the heavy hydrocarbons using a selec- tive precolumn. Or, you can distill the stock tank ol into gas and liquid components that {0 to the chromatograph and a heavy liquid residue that can be measured only for its molecular weight and specific gravity. ach method has its drawbacks. Injecting the petroleum liquid into the chromato- ‘raph risks polluting the column with heavy ‘elements that won't elute, which causes errors in the measured molecular weight ‘and specific gravity ofthe liquid cesiduo. Distillation takes 8 to 12 hours to perform, including the gas chromatography, and requires a minimum sample size of 40 cm Distillation can lead to inaccurate measure ‘ments of intermediate compounds (C; to Cx) that have a boiling point lower than ambient temperature. As the column reach: es ambient temperature at the end of the distillation, the intermediate compounds condense and fallback into the heavy liquid residue. The measured properties of the residue will be affected and the reported composition ofthe light distillation product will not account for the condensed interme- diate compounds. ‘At Schlumberger’s PVT lab in Melun, France, researchers Nikos Varotss and Paul Guieze have developed a fast and fully automated microdistllation process that replaces the standard distillation technique-* ‘This process separates mote efficiently the heavy residue, resulting in better composi- tional analysis. I takes one fourth the time to perform and requires only a few micro- liters (10 cm of sampling fluid. The molecular composition of a heavy stock- tank oil measured with microdistillation ‘methods compares well with that measured by conventional istllation(righd. The per- centage concentration of the residue is about three percent greater in distillation than in microdisillation. This difference indi cates the presence of intermediate com- pounds in the residue. In mictodistllation, a small sample is heated to a specific temperature while a ‘continuous inert gas flow removes the vapor proeiuced, which contains light compounds This continuous removal of vapor induces the heated sample to vaporize faster. The continuous stream of stripped com- pounds is carried by the inert gas into a The Technical ReviewNitrogen Detector Response ——> Cet) Normal Hexane (met) Propane CaHy) Ieopentane Isobutane “sio) ( Carbon Dioxide (Co,) = al | ee Eo & Isoheptane NI C,H) | Normal Pentane Cyt) Ethane (C,H) Normal Butane Typical tro: ‘matogram used for analyzing bydrcarbon ompostion. The ‘chromatogram pre- sents detector reponse versus retention te, the tm took a par ticular compound 0 ass through the Instrument’ adsorbing cokimns. Cheo- -matograms areca brated by injecting the ‘chromatograph with a ‘Sancard mature of| known compounds Normal Heptane eC He) Normal Octane CeHse) { Isooctane: ‘CaHi) Normal Nonane (-SgHog) \ | BF lecherane 64H, tiliation and Standard Distil ‘ouayolat Concentration % Microaistaton Distilation G 00 ‘oor Gr 008 one FO 0.06 0.08 ao, on oat 1 oz 030 705 a4 050 G 738) 205 G 876 aa G 1010 37 oy 398 o0T Ce 337 345 oe 666 Bat Cay 5.16 a0 Gr, Spee. avy 0.885 0.082 Volume 37 Number 1 a 700 low-temperature (less than -148°F [-100°C}) trap filled with adsorbent that Captures even the intermediate compounds ‘When the separation is complete, the trap is heated rapidly to desorb the trapped com- pounds and then flow them into the gas ‘chromatograph where they are identified with an FID, ‘Measuring the molecular mass of the residue and the liquid sample mass before and after the microdistllation yields the stock-tank liquid composition. The molecu- lar mass ofthe residue is measured using a gel permeation chromatograph (GPC), a type of chromatograph that uses a liquid solvent carrier, carbon tetrachloride, rather than a gas carrer’ {Yara Nae Giese Micodlaion A etd Combining Casa Liuid Chromatography to Cha ‘ze Pteleam Lid” ural of Como phy 03 09871 159170. 5 Gileze® ar Was IM: Detemiraon othe ‘Molar Ma of Patleu Diaon Reus Using, Ge Pemeation Cromatayaghy* ura of Cro ‘matgrophy 3121988): 261-272 "Reservoir Fingerprints ‘Compositional analysis techniques such as GPC allow a “fingerprint” analysis of both residues and lighter petroleum products. CCrudes from the same origin have nearly identical chemical fingerprints. Dissimilar fingerprints at different depths in the same ‘well suggest noncommunicating production zones (opposite, right). Dissimilar finger- prints from wells in the same field suggest that the wells produce from the different accumulations (opposite, below righd. When a mixture is injected into the GPC, molecules carried into the gel-filled columns are separated according to size. Smaller molecules representing the lighter ‘compounds follow the most tortuous paths through the gel and emerge last. Heavy compounds, which are too large to fit through the tiny interstices between gel par- ticles, emerge firs. An ultraviolet detector responds to the unsaturated bonds of arenes and an infrared detector responds to the alkyl fragments of alkanes (see “Composi tion of Hydrocarbons," below), The result ing distributions are plotted with a horizon- tal molar mass scale, rather than the time scale of a chromatogram. Equations of State During production the petroleum engineer ‘often needs information about reservoir fluid properties and phase behavior for pressure and temperature regimes not cov- ered during the initial PVT analysis. In the North Sea, for example, seawater injection used to drive production can lower the reservoir temperature some 10°F to 15°F {5°C to 8°]. Separator conditions, used to determine the formation volume factors and the solution gas-cl ratio, may also change. In these cases, petroleum engineers use an equation of state to simulate the phase behavior of reservoir fluids. familiar equ tion of state isthe real gas law, pV = ZnRT, in which Z is a dimensionless factor that Composition of Hydrocarbons ‘As the term suggests, hydrocarbons consist primarily of hydrogen, 11 to 14 percent by ‘weight, and carbon, 84 to 87 percent by ‘weight (ight), Oxygen, sulfur nitrogen, bel tum and metals exist as trace impurities ‘rude. The molecular structures of hydro- carbons differ greatly, however, and these sicuctures are classified into 18 different series, depending on whether carbon atom and hydrogen aioms are connected by open cchains or cyclic chains, and whether car- bon-carbon bonds are single (saturated) or ‘multiple (unsaturated), The amount of each series in the crude determines its properties Alkanes, eycloalkanes and arenes are the ‘most common hydrocarbon constituents i ‘rude ols. Alkanes, or parafin, are satura ‘ed hydrocarbons with carbon atoms arranged in open chains and single bonds (ar righd, The general formula is CyHzosa: ‘giving the first member of the series as ‘CH,-—methane, Two fragments ofthe alka- ne molecules, CH, or CHy, are called alkyl. ‘Alkanes are characteristcally stable, At stan- dard temperature and pressure, the frst four ‘members of the series—methane, ethane, propane and butane—are gases the next 13. members—CiH, to C,,Hyy-—are liquid and those from C,H. and above are solids forming parafin2 Ta nah enema Tate ‘pian She New Vor ah ley Sony 2, Buc rope of tly Orc Fela Daas n Roar etaeem Iesera fla, GowonIntemaona man Rescrces Development Coportion, 1978 2 Element Weight % Carbon a7 Hydrogen Ti4 Sutor 0.06-8.00 ‘Nivogen 002-470 ‘Oxygen 008-1.82 Wtals 0.00-0.14 Cycloalkanes are saturated hydrocarbons that form closed rings rather than chains (ight, middle), These follow the formula iH, the frst in the series being CsHe— cyclopropane. Like alkanes, they are very stable compounds, “The arenes, C,Han- also called aromat- ies for their fragrant odo, are an unsaturated series forming rings with double carbon bonds (righ, These compounds are les sta- bile than alkanes, but are not considered highly reactive, PVT analysts use a shorthand for describ- ing groups of chemical compounds. For example, the term C_ refers to all com- pounds with less than three carbon atoms; ,, stands forall hydrocarbons with four or more carbon atoms. Unless otherwise requested, a compositional analysis pro- Vides the mole fraction and percent weight ‘ofthe lighter compounds up o Cy, and the ‘mole fraction and percent weight of the remaining heavy compounds, collectively referred 10 a5 C2 accounts for the difference between real and ideal gas behavior, nis the number of ‘moles of a gas, and isthe universal gas constant, ‘Much more complicated equations of state have been developed that simulate the attractive and repulsive forces between hydrocarbon molecules—the most popular is due to D-¥. Peng and D.B. Robinson of, the University of Alberta, Edmonton, Cana- dda.* Such equations are based on the com- positional analysis of the mixture and PVT relations derived experimentally. An equa: tion of state must be calibrated, or tuned, against real PVT data. This involves adjust: ing parameters inthe equation until the pre- dicted and actual PVT data agree. Fluid Analysis at the Well ite ‘A thorough PVT analysis takes about three ‘weeks to process and in the case of remote field locations, expedition and transport continued on page 14 sta ih ig Propane (C4Hg) Cycloalkane Series oH, on ch, CHy—CH, ‘Gyelopentane (CH) ‘Avene Series 4 | Hl ve a Ce. ae ie —) H Berwene (C,H) Lsructre of rpresertative members ofthe tree most common consents of crude s—the alkane, eyeloatane, and arene series, Propane (CsH9 represents the alkane series which are sraight chain or branched chain “arrangements of hyragen and carbon atoms (Gyclopentane (CH) represents the cyclolkane ‘sees which are strated hydrocarbons forming Closed rings. Boze (CoH. represenis tho ‘arene series, unsaturated hycrocarbons that con- fain double carson bond.Lower Zone Upper Zone Molecular Weight = 194 Molecular Weight = 145 Normalized Distribution Utaviolet Response (nsaturates) Utraviolet Response (Unsaturates) Le 100 00 700 000 c cn °, Cn Molar Mass, g/mole Molar Mass, g/molo {Ae permeation chromatography (GPC) fingerprints taken 15 feet (5 meters apart in depth in a US Gulf Coast wel, The dissimilar alkyl responses indicate that these uid originate fom noncommunicaing production zones. Well A Well B Molecular Woight = 146 3 é g lettered i Fesponee—| 5 ‘Algis) 2 Utvaviolt Response (Unsaturates) fy “o 700 0 i ‘000 c Ce ©, ow Molar Mass, g/mole Molar Mass, g/mole Gel permeation chromatography (GPC) fingerprints taken fom two wells inthe same fed. The similar alkyl responses reve! thatthe wll produce ‘rom separate reservois. “7 Reng 0 ac Raina DB: "A New Two Ceraat quston St" IndatalEroering Chom, ander 15 0976) 59.68, For atic view ofthe evltion of equation of ‘ie fom vn der Waist present ‘onepados Cand Heda "Fem Reich Kwong oe Pes” Mud Phase Eu (0368123. Volume 37 Number 1 13tion problems can sometimes delay the results for months. Inthe meantime, crucial reservoir engineering decisions are based on fluid properties derived from emprical cor- relations that relate easily obtained reservoir properties to parameter like the ol forma- tion volume factor, the bubllepoint, and the dewpoint. ‘A widely used correlation developed by MW, Standing of Standard Oil Co., La Habra, California obtains the bubblepoint fom cil and gas gravities, reservoir tempera ture and the gas-oil ratio (GOR)? Standing’ correlation is based on 105 data points from 22 different California crude samples. ‘Another bubblepoint conelation, developed by ©. Glaso of the University of Trondheim, Norway, is based on North Sea samples.® Correlations are appropriate, however, only when the reservoir sample is similar to the samples that generated the correlation, ‘A new Flopetrol-Schlumberger service called the Fluid Properties Estimation (FPE*) system, also developed by Varotsis and Guieze, reduces the need for correlations in the field by providing on-site measurements ‘f fluid properties and compositional analy- sis that are both used to tune an equation of state.? The equation of state can then be used immediately to predict the phase bbchavior and PVT properties of the fluids to be produced at reservoit, well and surface conditions. ‘The FPE approach does not replace a ‘complete PVT analysis, but rather allows a ‘proper validation of samples on site—mak- ing sure that the sample is representative of reservoir fluid above its bubblepoint. It also provides a “quicklook” fluid analysis that is especially needed infield locations without ready access toa PVT lab. For compositional analysis, the FPE sys- tem consists of two portable gas chro- ‘matographs and a personal computer that controls the chromatography, acquires and processes the output signal from the chro- ‘matograph and performs the PVT simulation f the oil with the equation of state (far righ Other measurements performed on the samples include the bubblepoint speci ic gravity, viscosity and refractive index. Currently, the analysis works best with bottomhole samples (see “Sampling Sum- ae ; ee eee E mary,” opposite). Surface samples can be used, but the bubblepoint pressure will not be measurable. A bubblepoint estimate ‘must be made using correlations, and this could affect the accuracy of FPE predictions. The chromatograph dedicated to gas analy- sis contains two packed columns, each with own TCD and a column switching device (above). The injected gas enters the first column, and, after a few minutes, the light compounds including methane, nitro- ‘gen and oxygen elute out of the column together. This group of compounds pro- ‘ceeds to the second column where it elutes as individual peaks. By this time, com- pounds C; through C; which continued advancing through the first column are being separated and detected. The heavier ‘compounds (C,,) are backflushed out of the first column and detected as one lump peak by the first TCD. The chromatograph dedicated to liquid analysis contains a packed precolumn, a thin column coated on the inside with an absorbing liquid and a TCD, Stock-tank oil is placed in the injector and partially vapor- ized. The distillate proceeds to the precol- ‘umn where itis split into a light part (C.) and a heavy pat (Cy). The light part is sent to the column and the heavy part is back- flushed out of the precolumn. This prevents ‘contamination ofthe column withthe heavy ‘compounds. TAL aun 2 (the gas ehromato- ‘gaph ofthe Fld Properties Estimation PER system dec. ed '0 anayts of gas samples. Is two Colum: can be con- ected in Sries orn parallel. (a) The sam. le is nitally injected Swi the columns fn Series. (b)When the fist thermal eondc- tity detector (TCD) ‘tects the methane ‘pel the computor Switches the columns topparae allowing oxygen and nitogen, ‘which re tially detected together 0 be separately ident fied bythe second TCD.(@) When the fist TCD detetsC,, the computer reverses the camer flow to backfethe Coy ‘compounds out othe fist column deci it the ist TCD. actor 2 Cy_ Soparatod ihio Oxygen. Nivogen, act Methane Peaks (Othe Fluid Properties Estimation (FPE System used for onsite PVT analysis. The ruggedized FPE system houses tw portable gas ehro- ‘matographs anda personal computer that con luoks the chromatographic uns, acquit and pro- ‘cesses the ouput signal rom the chromatograph ‘and performs PVT simulations based on ae tion of state 7, Sandi M.7A Presi lane Tergeatur Cor teat er Mitre of Caloris Os rd Cass Dulin and radi Paces New York. es ‘can Pele iste (19471 275207, 18. Chen ©: General rere Velue- Tange {ure Corehtions” Jounal of Peoleum fechaogy 52 980745 795. 9. Vols Nand Gulez Onsite Reser Fa “Arla” paper SPE 1817, psa tthe SPE (int Anna echnical Conierence a Ebon, Houston, ether 25,1988. + Mark Schiumbee The Technical ReviewSampling Summary Stook-Tank Oi Gi Sample {Sure saming of and 2s and as saps collec inatngcsl forthe spat (o ensure identical Sampling conditions and appropriately ecombinable actions. The temperate {nd presi othe separator and stock ane are recorded dn sampling Successful PVT analysis requires that fluid ‘samples represent the original hydrocarbon, Inthe reservoir. Unlike coring and logging programs, which are performed routinely throughout reservoir development, fluid sampling for PVT analysis must occur soon. after the exploration wells are drilled in ‘order to characterize the original reservoir fluid, Once the reservoir pressure falls below the bubblepoint, gas comes out of solution and will accumulate to a critical saturation in the rock pores before flowing to the well, Similarly In a gas condensate well, cil begins to condense once the pres- sure falls below the dewpoint. The oil in a as condensate, moreover, may never reach A critical saturation and flow. In elther case, the composition of the fluid entering the ‘wellbore differs ftom its original sate. ‘Sampling procedures depend on whether the reservoir Is saturated (its pressure equals the bubblepoint) or undersaturate (its pres: sure fs above the bubblepoint).! Before sam= pling the well, reservoir engineers use core lations to estimate the bubblepoint from ‘well test data such as the initial and present reservoir pressures, reservoir temperature, oil and gas gravities, and stabilized produce tion GOR at surface. Most wells must be conditioned before sampling because the initial pressure draw- down with production alters the nature of the fluids around the wellbore. The goal of Volume 37 Number 1 conditioning is to ensure that fluid entering the wellbore is identical to the original reservoir fluid, This is accomplished by pro- ducing the well at successively lower rates Lunt the GOR stabilizes. Reservoir fluids can be sampled down- hole or at the surface, In surface sampling, samples of separator gas and oil are collect: ‘ed simultaneously along with rate measure= ‘ments and recombined to form a reservoir fluid sample (above), Between two and four 10- oF 20st samples of gas, depending on the GOR, and 600 cm? of oll are collected Downhole sampling employs a sampler tool that is run with a pressure gauge on either wireline or slckline (righ). After low= ring the sampler to the top of the produc- ing zone, the sample chamber is opened by a clack (for sliekline tools) or an electric si nal from surface (for wireline too!s). The oil flows slowly into the tool at constant pres- sure to avoid gas coming out of solution, A piston seals the chamber and the tool is brought to surface. A minimum of three samples should be collected to ensure that atleast one is valid “The advantage of surface sampling is that iteasly provides large fluid samples, but the GOR used to recombine the oil and gas samples must be accurately measured, Downhole samples are validated intially in. the field and later in the PVT lab by measur- ing the sample's bubblepoint and checking that it Is less than the sampling pressure. ‘Another check is to compare the bubble- points of the three or so collected sam- ples—if they agree then the samples are Considered representative. If the measured bubblepoint exceeds the downhole sam- pling pressure, then the sample container has leaked oil or collected fee gas. - oS ‘vol has been propery coneliloned, the sarpler Is lowered tthe desired depth and activated to Collect a sample. Once filed, a piston seas the ‘chamber and the tools raeved. At eas! tree botomote samples should be collected 1, Guidelines for Reser Fluid Smolin Meln, France opel Joboton, 1973, ‘AP Recommend Practice for Sang Peteum ‘Reserva Pid Dalles: Avetcan Pee is 15