ROCK AND FLUID SAMPLING AND ANALYSIS

Fluid Sampling & Analysis

Reservoir Fluids: Overview

Reservoir Fluids

Petroleum accumulations are defined as naturally occurring mixtures of organic compounds, primarily hydrocarbons, that are found within the earth. Beyond this shared definition, specimens of petroleum collected from various locations might appear to have very little in common. Petroleum samples can vary from clear liquids that look much like water to thicker, greenish or reddish-brown mixtures, or even highly viscous, semi-solid black substances. The odor of these hydrocarbons can range from a sweet, aromatic bouquet to the distinctly disagreeable smell of rotten eggs. Although the bulk of the chemical compounds that comprise crude oil are hydrocarbons, the fact that carbon atoms have the ability to form long branching and cyclic chains allows an almost limitless diversity in the molecular composition of petroleum accumulations.

Different specialists involved in petroleum exploration and production have different reasons for wishing to examine and characterize the hydrocarbon fluids and water found together in petroleum reservoirs. For example, a chemical engineer may be interested in a crude oil's composition only as it relates to the amount of commercial products the oil will yield after refining. An exploration geochemist might have an interest in an oil or reservoir water's composition, insofar as it sheds light on the origin, maturation, and degradation of the oil or helps point the way toward a better geological interpretation. The petroleum engineer is particularly concerned with the analysis of hydrocarbons in order to determine their behavior under the varying conditions of pressure and temperature that occur in the reservoir and piping systems during the production process.

Chemistry of Hydrocarbons

Despite the wide differences in the appearance of crude oils obtained from various localities, an ultimate analysis of these crudes for their weight percent of carbon, hydrogen, sulfur, nitrogen, and oxygen will show only minor differences. Table 1, below, gives the limits of these elements for nearly all crude oil samples. This phenomenon of consistency is due to the fact that hydrocarbon molecules have specific ratios of hydrogen and carbon atoms. This means that all hydrocarbons belong to one of a few series, although each series may have many thousand members. The differences in the properties of crude oils, therefore, may be explained by the relative amounts of each series present in the crude, although the relative amounts of carbon and hydrogen are not too different for each series. Hydrocarbons are classified as belonging to a particular series based on their molecular structure. Specifically, classification follows the arrangement of the carbon atoms (straight chains, branched chains, or cyclical), and the number of bonds between the carbon atoms (single, double, or triple).

Element	Weight %
Carbon	84-87
Hydrogen	11 - 14
Sulphur	0.06 - 8.00
Nitrogen	0.02 - 1.70
Oxygen	0.08 - 1.82
Metals	0.00 - 0.14

Table 1.
Element composition of crude oils (after Banks and King, 1984. courtesy of John Wiley and Sons)

Hydrocarbons that are members of the paraffin or alkane series are common constituents of many crude oils. Alkane hydrocarbons are straight-chain or branched-chain configurations of carbon and hydrogen atoms that follow the general formula, C_nH_2n+2. The first member of the series (methane) has a formula CH₄, the second member (ethane) is C₂H₆, and so on. The alkanes are saturated, that is, the carbon atoms are connected with single bonds. Figure 1 (Models for methane and propane, showing the tetrahedral nature of the carbon-hydrogen configuration, after Burcik, 1979) shows the models used to visualize the structure of these hydrocarbons, and their "shorthand" formulas. As longer chains are built, it becomes possible to arrange the carbon atoms in either linear or branched fashion without changing the relative number of carbon and hydrogen atoms. These different arrangements are called isomers and possess different physical properties. For example, Table 2, below, gives the formulas and several physical properties for normal pentane, isopentane, and neopentane. Obviously, as the number of carbon atoms in a molecule increases, so does the number of possible isomers.

Formula	Name	Molecular weight	Boiling point(°F,14.7 psia)	Critical press. (psia)	Critical temp. (°F)
CH₃ CH₂CH₂ CH₂ CH₃	n-Pentane	72.151	96.92	488.6	385.7
	Isopentane	72.151	82.12	490.4	369.03
	Neopentane	72.151	49.10	464.0	321.08

Table 2.
Isomers of pentane showing the slight changes in physical constants due to the differences in structural formulas (from Engineering Data Book, 1972; reprinted by permission of NGP5A)

All the straight-chain alkanes from CH₄ (methane) to C₄0H₈₂
(tetracontane) have been identified in crude oil. Typically, they amount to 15% to 20% of the oil. The possible isomers for these alkanes range from two for butane to about 6.2 10¹³ for tetracontane (Banks and King 1984). The alkanes are characterized by their chemical inertness, which probably accounts for their stability over long periods of geologic time. The first four members of this series (methane, ethane, propane, butane) exist as gases under standard conditions of pressure and temperature. Those from C₅H₁₂ (pentane) to about C₁₇H₃₆
are liquids, and C₁₈H₃₈ and higher are waxlike solids. "Paraffin" is a mixture of these solid members of the series.

Saturated hydrocarbons that form closed rings rather than chains belong to the series known as cycloalkanes (also called naphthenes or cyclopraffins). These hydrocarbons follow the general formula CnH2n. Being saturated, these hydrocarbon compounds are relatively stable and possess chemical properties similar to those of the alkanes. Two very stable members of this series, cyclopentane and cyclohexane are shown in Figure 2 (Structure of several members of the cycloalkane series of hydrocarbons).

Unsaturated hydrocarbons are compounds that contain a carbon-carbon double bond. That is, two valence electrons of each of two carbon atoms are involved in bonding the two carbons together.

These compounds can add hydrogen to their structures under appropriate conditions, and are therefore said to be unsaturated (with hydrogen).

One class of hydrocarbons that contains carbon-carbon double bonds is the arene (also called aromatic because many of them have fragrant odors) series. This group is made up of derivatives of benzene, whose formula is C6H6, and whose structure is shown in Figure 3 (Structure of several members of the arene, or aromatic, series. The formula convention used here to represent benzene is called a Kekule structure. The compounds shown are common constituents of crude oil.)

Although the benzene ring contains three double bonds, its unique structure allows it to be relatively stable and unreactive. The extra pairs of electrons are part of the overall ring structure and are not localized to one pair of carbon atoms. Some of the simpler members of this series consist of a benzene ring with one or more alkyl (CH₃) groups attached as side chains in place of hydrogens. These compounds are common constituents of crude oil and include toluene and xylene. The arene hydrocarbons are either liquids or solids under standard conditions.

Other unsaturated compounds are called alkenes or cycloalkenes, depending on their straight-chain or cyclic structures. They are known as olefins. Another series of unsaturated hydrocarbons is the acetylene series, which includes compounds having triple carbon-carbon bonds. Olefin compounds are very uncommon in crude oils and the acetylene series is virtually absent. This is undoubtedly due to the high degree of reactivity of these compounds, and their tendency to become saturated with hydrogen, forming alkanes (Banks and King 1984).

Of the eighteen different possible hydrocarbon series, therefore, alkanes, cycloalkanes, and arenes are the common hydrocarbon constituents of most crude oils. One historical approach to the characterization of crude oils has been to classify them by base as either paraffinic, naphthenic, or intermediate. The term "naphthenic" is misleading here, for it is not directly related to the hydrocarbon series mentioned earlier, but simply indicates an oil that, upon distillation, forms a residue of asphaltic rather than waxy material (Banks and King 1984).

The U.S. Bureau of Mines developed a classification system in which a crude sample is distilled and the key fractions recovered between specific temperatures tested for their specific gravity. This approach produces a range of nine possible classifications. These groups can be further subdivided by the determination of the waxforming characteristics of the second (high boiling point) fraction.

Crude oil	Base	API gravity of key fraction 1	API gravity of key fraction 2	UOP factor (average)
Pennsylvania, USA	Paraffinic	40 or lighter	30 or lighter	12.2 - 12.5
Mid Continent USA	Intermediate	33 - 40	20 - 30	11.8 - 12.0
Gulf Coast, USA	Naphthenic	33 or heavier	20 or heavier	11.0 - 11.8
Leduc, CAN	Intermediate	33 - 40	20 - 30	11.84
Lloydminster, CAN	Naphthenic	33+	20+	11.35
Kuwait	Paraffinic-intermediate	40 or lighter	20 - 30	11.93
Abqaiq,	Paraffinic-	40 or lighter	20 - 30	12.00
SAU.AR.	intermediate
North Sea	Intermediate	33-40	20-30	11.84-11.98

Table 3. Classification of crude oil by base and UOP characterization factors. High UOP factors indicate more paraffinic composition, low UOP factors indicate a greater percentage of naphthene and aromatic compounds (after Buthod, 1962; courtesy of SPE/AIME)

An attempt to establish a single index number that would provide a correlation for the base of a crude oil resulted in the U.O.P. characterization factor (Universal Oil Products Co.). This number is based on the average boiling point and specific gravity of a sample: values between 12.5 and 13.0 are found for paraffinic hydrocarbons; between 11.0 and 12.0 for naphfhenic hydrocarbons; and between 9.0 and 12.0 for aromatics. Table 3, above, shows the base and U.O.P. factors for several typical crude oils. These classification systems are mentioned here primarily because of their wide coverage, and because practically all known crude oils have been analyzed and classified by these methods and references to them are still common. In general, these classification methods serve as a guide to the ultimate commercial value of refined products obtained from a given crude.

Modern crude oil classification commonly is based on a division of the hydrocarbons present into three main groups: alkanes, naphthenes, and aromatics, in combination with nitrogen, sulfur, and oxygen compounds (called resins and asphaltenes). This system can be depicted with a ternary diagram, as shown in Figure 4 (Ternary diagram showing the composition of the six classes of crude oils according to Tissot and Welte. Samples are shown from 541 oil fields-- from Tissot and Welte, 1978, reprinted by permission of Springer-Verlag). Here we see 541 oils plotted according to the percentage of these three groups in their composition. There are few oils with more than 50% naphthenic content. Paraffinic-naphthenic, paraffinic, and aromatic-intermediate oils are most common (Tissot and Welte 1978).

Natural gas is composed primarily of the lighter (low carbon) members of the paraffin series. Methane and ethane frequently comprise 80% to 90% by volume of a natural gas (Amyx, Bass, and Whiting 1960). A natural gas with a greater percentage of high carbon constituents will typically yield larger amounts of natural gasoline or other liquids.

All hydrocarbon mixtures, whether natural gas or crude oil, contain various amounts of nonhydrocarbon substances. Crude oils typically contain organic compounds which incorporate various combinations of sulfur, nitrogen, and oxygen atoms. Some organic compounds also contain metallic elements, particularly vanadium and nickel. While these components are distributed throughout the whole distillation range of a crude oil, they tend to be concentrated in the heavier components. The color and odor of crude oil stems mainly from the nitrogen, sulfur, and oxygen compounds concentrated in the C₂₆ — C₄₀+ fractions (Banks and King 1984). Sulfur is the most common nonhydrocarbon constituent, and the average sulfur content of crude oil is 0.65% by weight (Banks and King 1984). Carbon dioxide, nitrogen, hydrogen sulfide, helium, and hydrogen are the more common impurities found in natural gas hydrocarbon mixtures. For our purposes, these impurities are important because they can cause deviations from typical hydrocarbon behavior if they are present in sufficient quantities. In addition, impurities such as hydrogen sulfide and carbon dioxide can also cause severe corrosion and handling problems for a production engineer. Sulfur contamination can also lead to serious refining problems for chemical and process engineers.

Detailed analytical separation of crude oil or natural gas samples into components far beyond C₆H₁₄ (the hexanes) is expensive and not commonly done for most oilfield applications. As we shall see in later sections of this module, laboratory measurements of physical behavior or correlations based on easily measured properties are sufficient for most petroleum engineering purposes.

Classification of Reservoir Fluids

Common oilfield classifications of oil and natural gas rely on observed producing characteristics and easily measured specific gravity. The gas-oil ratio (GOR), gas gravity, and oil gravity are used to categorize reservoir hydrocarbons. Gas-oil ratio in this case refers to the cubic feet of gas produced per barrel of liquid (or cubic meters per cubic meter), with both volumes measured at standard conditions of temperature and pressure. As we shall see, the pressure and temperature of surface separators, and the number of separation stages, control the gas-oil ratio for a given system.

Specific gravity, of course, is the ratio of the density of a substance to the density of some reference substance. For gases, the standard reference is dry air at the same temperature and pressure as the gas in question. These conditions should be specified. Specific gravity is also equal to the ratio of the molecular weight of the gas to the molecular weight of air (28.97). For liquids, the reference is pure water at 60 °F (289 K) and one atmosphere (14.7 psia or 101 kPa).

For hydrocarbon liquids, the API gravity (American Petroleum Institute) scale is most commonly used in the oil industry. A similar scale is the European Baume gravity scale. These two oil gravity scales expand and invert e range of numerical values for oil specific gravity. For example, a very heavy oil might have a specific gravity of 0.96, while a light distillate might be 0.78. Expressed in degrees API, these same gravities would range from 160 (heavy) to 500 (light). Baume gravity values are very close to API gravity over the general oil range. These scales allow gravity measuring hydrometers to be easily calibrated with a linear scale, and values are often correlated with other oil properties. Water has an API and Baume gravity of 10.0. The relationship between API gravity and specific gravity is given as:

Reservoir fluid	Surface appearance	GOR range	Gas specific gravity	API gravity
Dry gas	colorless gas	Essentially	0.60 -0.65 no liquids
Wet gas	Colorless gas with small amount of clear or straw- colored liquid	Greater than 100 MSCF/bbl	0.65 - 0.85	60° - 70°
Condensate	Colorless gas with significant amounts of light- colored liquid	3 to 100 MSCF/bbl (900-18000 m³/m3)	0.65 - 0.85	50° - 70°
"Volatile" or high shrinkage oil	Brown liquid with various yellow, red, or green hues	About 3000 SCF/bbl (500 m3/m3)	0.65-0.85	40° -50°
"Black" or low shrinkage oil	Dark brown to black viscous liquid	100 - 2500 SCF/bbl	(20 - 450 m3m3)	30° - 40°
Heavy oil	Black, very viscous liquid	Essentially no gas in solution		10° - 25°
Tar	Black substance	Viscosity	>10,000 cp	<10°

Typical composition, mole %
	C1	C2	C3	C4	C5	C6+
Dry gas	96	2.7	0.3	0.5	0.1	.4
Wet gas
Condensate	87	4.4	2.3	1.7	.8	3.8
"Volatile" or	64	7.5	4.7	4.1	3.0	16.7
high shrinkage
oil
"Black" or low	49	2.8	1.9	1.6	1.2	43.5
shrinkage oil
Heavy oil	20	3.0	2.0	2.0	2.0	71
Tar	—	—	—	—	—	90+

Table 1.
General categories of reservoir hydrocarbons. There are no definite boundaries between these classifications and usage may vary depending on location. Gravities and GOR are also dependent on separation conditions

Table 1. shows the classifications of reservoir fluids, based on GOR and fluid gravities. A typical composition is also given, showing the mole percent of the lighter paraffin constituents relative to the heavier hydrocarbons. Remember that a mole is one molecular weight unit of any substance. Thus, a pound-mole is an amount of a substance equal to ifs molecular weight in pounds. The molecular weight of methane is 16.04, and one pound-mole of methane is equivalent to 16.04 pounds of methane. Similarly, one gram-mole is 16.04 grams of methane.

It so happens that in the case of ideal gases, moles of any two gases occupy equivalent volumes and contain equivalent numbers of molecules at equivalent temperatures and pressures. So in this sense a mole can be thought of as a discrete number of molecules. For a mixture, the mole percent of a component in the total volume is the fraction of molecules in the mixture that are that component.

The categories in Table 1 are not rigorously defined. In general, low GOR, low API-gravity oils have lesser amounts of the light paraffinic hydrocarbons, while dry gases are composed almost entirely of these compounds. As we shall see, sampling and analyzing the behavior of dry gas or black oil systems are relatively straightforward procedures. Condensate and volatile oil systems, on the other hand, can be much more complex in terms of their physical chemistry.

Characteristics of Formation Water

In addition to liquid and gaseous hydrocarbons, petroleum reservoirs always contain a third fluid, water. Formation water is termed connate water if it is believed to be a remnant of the original water in which the sediment was deposited. Meteoric water refers to formation waters that originate as rainfall and are carried into the ground via outcrops, fractures, or permeable sediments. Interstitial water is the preferred term for the formation water that shares the pore space of the hydrocarbon reservoir with oil and gas, regardless of origin. Interstitial water saturations in petroleum reservoirs usually range from 10% to 50% of the pore space, and water saturations can vary throughout a reservoir, depending on the pore structure of the rock (Levorsen 1967).

Formation waters are most commonly distinguished by their varying degrees of salinity, that is, the amount of dissolved ions present in the water. The ionic composition of formation water is usually measured in milligrams per liter and expressed in milligrams per liter (mg/I) or parts per million (ppm).

A unit volume of solution is taken to have a million parts of weight, and the number of parts of weight contributed to the solution by a given ionic component is the parts by weight per million. Parts per million is only equivalent to milligrams per liter at a solution density equal to that of pure water. Since most formation waters have densities close to this value, the difference between numerical values in mg/I and ppm is not great. However, for an extremely salty water the difference could be as great as 20%.

Composition, parts per million
Common ions	Chemical name	Seawater	Cretaceous,sandstone Burgan Field Kuwait	Miocene formation Lagunilla Fields Venezuela	L. Cretaceous limestone Rodessa Field, Texas-La.
Cations
Na+ and K+	Sodium and potassium	11,000	46,191	2,003	61,538
Ca++	Calcium	420	10,158	10	20,917
Mg++	Magnesium	1,300	2,206	63	2,874
Anions
Cl-	Chloride	19,350	95,275	89	140,063
SO4=	Sulfate	2,690	198	—	284
CO3=	Carbonate	150	—	120
HCO3-	Bicarbonate	—	360	5,263	73
	Total ppm	34,910	154,388	7,548	225,749

Other common ions include: barium (Ba++), lithium (Li++), iron (Fe+), Nitrate

(NO3- ), bromide (Br-), iodide (I-), sulfide (S=) and strontium (Sr++).

Table 5. Composition of typical formation waters. The composition of seawater is given for comparison (after Levorsen, 1987; courtesy of W.H. Freeman Co.)

Table 5 shows some typical formation water compositional analyses. The total of all ionic components is often termed "total dissolved solids," and used as a reference index in comparison of formation waters. More informative techniques for categorizing formation waters involve the plotting of ionic composition on diagrams, as shown in
Figure 1 (Formation water composition diagrams. Shown here are the (a) Tickell method and the (b) Stiff method for seawater). These graphs are useful in making quick visual comparisons of formation water from different formations.

The Tickell method (Tickell 1921) uses the reaction value of the various ions as a value for graphing. The reaction value is equal to the mg/l concentration times the valence, divided by the atomic weight. Thus 420 mg/l calcium is equal to a reaction value of 420 2 ÷ 40.08 = 20.9. If the reaction values for all ions are added together, a particular ion's reaction value may be expressed as a percentage of the total. Reaction values are also called equivalents per million, or milliequivalents per liter. The Stiff method (Stiff 1951) also uses milliequivalents per liter, but the scale is horizontal, resulting in a "butterfly" pattern. Usually the sodium and chloride scales are graduated in scale units of 100 milliequivalents, and the other ionic scales in units of 10, although the scales can be varied for any group of diagrams under comparison.

Salinity is the most important characteristic that can be measured for a formation water sample. The ways in which formation water and hydrocarbon mixtures change with pressure and temperature depend upon the degree of salinity. For example, the solubility of natural gas in a formation water with 150,000 ppm total dissolved solids is only about half the solubility in pure water. However, even in pure water, the solubility of natural gas is not great; it is only about 10 to 30 SCF/ bbl (2 to 5 m³/m³) under most reservoir conditions.

Determination of the formation water salinity is more important for correct interpretation of electric well logs. The resistivity of a formation water is related to its salinity, and a value of formation water resistivity is necessary for the quantitative evaluation of resistivity logs. Although this value can be estimated from log data and correlations, the best results are obtained when a sample of formation water is retrieved and analyzed.

Applications of Reservoir Fluid Data

Before we discuss the behavior of reservoir fluids, and the techniques for analyzing that behavior, we should briefly review our reasons for acquiring this information. As we shall see, there are a variety of correlations available for estimating fluid properties based on a few easily obtained measurements. These methods are, of course, somewhat less precise than a complete laboratory analysis of a reservoir fluid. However, correlations are valuable aids in the design of our ultimate sampling and analysis procedure, and are often the only data available for preliminary calculations, proposals, and estimates.

Figure 1 (Applications of reservoir fluid property data in different areas) is a generalized matrix showing the application of fluid property estimates and analyses in different areas of the exploration and production process. While reservoir engineers generally have the greatest claim on such data, reservoir fluid analyses are also quite valuable to geologists and production specialists.

Often the process of applying fluid data is one of revision and improvement. For example, a geologist may use correlations along with an oil or gas gravity measurement from a nearby well for help in obtaining an estimate of the potential reserves to be found in an exploration prospect. After the exploration well is drilled (and successfull), a well test may allow those same correlations to be used with the known gravity, GOR, and pressure data from the discovery well. In an ideal situation, a fluid sample may be recovered from the discovery well for analysis. This more precise information on the properties of the hydrocarbon accumulation may be used by geologists and engineers to justify further development drilling. One or several of the development wells may then be completed and reservoir fluid samples retrieved. The laboratory analysis of such samples provides the more accurate information needed to help plan the development of the field, design production facilities, determine the size and cost of equipment, and thereby make economic decisions. After production has been established, further sampling and analysis may be requested by the engineer to evaluate potential improved recovery projects. If such a project is implemented, routine sampling may even become a necessary part of the project's maintenance.

All of this points toward a general increase in the amount of reservoir fluid sampling and in the accuracy of the overall interpretation of reservoir behavior, as the development of a field or reservoir progresses. However, the degree of accuracy required in engineering calculations is usually directly related to the cost of the proposed project. In the past, unless an expensive enhanced recovery project was being considered, the cost of obtaining good reservoir fluid samples for reservoir analysis was often difficult to justify. However, the costs of development in today's industry, particularly in the offshore areas, are making early routine fluid sampling much more commonplace.

The value of the data obtained from a sampling procedure, however, is only as good as the procedure itself. It is vitally important that the proper procedure be followed in obtaining a reservoir fluid sample, or the data will be no better than that estimated from correlations, and perhaps worse (and a good deal more expensive).

Exercise 1.

You receive a production report that includes the following information:

Total oil rate = 1250 STB/day (198.7 m³/day)

Total gas rate = 3565 MSCFD (100,950 sm3/day)

Oil gravity = 35 °API at 14.7 psia and 60 °F

Gas gravity = 0.656 at 14.7 psia and 60 °F

Calculate the gas-oil ratio, the oil-specific gravity, and the molecular weight of the gas.

What general category of reservoir fluid does this appear to be?

Solution 1:

The gas-oil ratio is calculated as:

= 2852 SCF/STB

or
=508 sm³/m³

The oil-specific gravity is calculated from the equation

where:

Specific gravity =

= 0.85

Remembering that the specific gravity of a gas measured at standard conditions is the ratio of its molecular weight to that of air:

MW = (0.656)(28.97) = 19.00

This appears to be a reservoir fluid in the lower end of the volatile oil range, or the upper end of the "black" oil category — only a reservoir fluid analysis can precisely predict its behavior.

Exercise 2.

The following producing characteristics were observed for specific reservoirs. How would you categorize the reservoir fluids?

Sajaa Field, Sharjah UAE

GOR = 116 STB/ MMSCF

Gas gravity = 0.81

Oil gravity = 51.3° API

Composition, mole %

C1 = 76

C2 = 5.6

C3 = 2.9

C4 = 2.2

C5 = 0.6

C6+ = 6.6

N2 = 0.7

CO2 = 4.6

H2S = 0.1

Eugene Island, Offshore LA

GOR = 275 SCF/STB

Gas gravity = 0.61

Oil gravity = 34 °API

Composition, mole %

C1 = 39

C2 = 1.7

C3 = 2.2

C4 = 2.8

C5 = 2.9

C6+ = 51.6

N2 = 0.01

CO2 = 0.16

H2S =

Hugoton Field, Oklahoma & Texas

GOR =

Gas gravity = 0.71

Oil gravity =

Composition, mole %

C1 = 72

C2 = 7.0

C3 = 7.7

C4 = 0.3

C5 = 0.7

C6+ =

N2 = 15.5

CO2 =

H2S =

Sintef Field

GOR = 648 SCF/STB

Gas gravity = 0.695

Oil gravity = 34 °API

Composition, mole %

C1 = 46

C2 = 3.7

C3 = 2.0

C4 = 1.5

C5 = 0.8

C6+ = 45.4

N2 =

CO2 =

H2S =

Solution 2:

The Sajaa reservoir fluid is identified as a gas condensate. This is apparent from the relatively rich liquid content of the gas (116 STB/MMSCF = 8620 SCF/STB and 652.6m³/ mmsm³ - 1532 sm³/m³) and the high oil gravity. An analysis of this reservoir fluid indicates that although the original fluid exists as a gas, with pressure decline in the reservoir a liquid phase will form.

The Eugene Island reservoir fluid is typical of many offshore Gulf Coast fields and is a black oil. The lack of hydrogen sulfide means that this oil is relatively easy to transport, store, and refine.

The Hugoton gas is from a well-known dry gas field. The high gas gravity is not a result of heavy hydrocarbons, but is due to a fairly high percentage of nitrogen, which has a molecular weight of 28.013 compared to methane at 16.043.

The Sintef fluid is a black oil.

A.2. Hydrocarbon Phase Behaviour

Phase Behavior and Physical Properties of Reservoir Fluids

The pressure and temperature conditions acting on a substance determine whether it exists as a solid, liquid, gas, or some combination of these phases. We are familiar, for example, with the condensation of wafer vapor by the lowering of ifs temperature. One can observe this phenomenon readily if one is wearing glasses when walking out of an air-conditioned building on a hot, humid day. Other everyday examples are butane lighters and propane gas stoves, which operate by the vaporization of liquid hydrocarbons with a drop in pressure.

Reservoir hydrocarbons will also change phase when subjected to changes in pressure and/or temperature. The precise manner in which these phase changes occur, that is, the phase behavior of a hydrocarbon mixture, will depend on the relative amounts of individual hydrocarbon compounds present. This is because the factors governing the phase behavior of hydrocarbons include the molecular forces of attraction and repulsion. For example, increased pressure and the attractive forces between molecules tend to move molecules together, increasing the density of the material, as when a gas becomes a liquid. This attractive force increases until the molecules get so close together that their electrical fields overlap, and then a repelling force tends to move the molecules apart. Increasing the kinetic energy of the molecules by increasing the temperature, will also tend to force the molecules apart.

As a result, as pressure is increased, molecules will be forced together and a gas will become a liquid. This is easier to accomplish when the molecules have a greater degree of attraction for one another, as in the case of the heavier, slower molecules. The attractive forces of these molecules are less likely to be disrupted by the speed at which the molecules collide or come close to each other. The attractive forces of lighter molecules, are more likely to be disrupted by the greater speed of these molecules, and therefore these molecules are more likely to be separated by increases in their kinetic energy (temperature increases). Thus, we have lighter, methane molecules easily forming a gas phase and heavier, heptane molecules forming a liquid phase under the same conditions of standard temperature and pressure. Of course, at higher temperatures both may exist as gases or at higher pressures and lower temperatures both may exist as liquids. A quantity of hydrocarbon molecules may even exist as two phases with a portion in the gas phase and a portion in the liquid phase. The relative amount of molecules in each phase is dependent on the balance of the forces tending to confine the molecules (increased pressure, attraction) and those tending to separate the molecules (increased temperature, repulsion). When all these forces are balanced, the phases are considered to be in a state of equilibrium.

Hydrocarbon Phase Behavior

It is customary to represent the phase behavior of hydrocarbon fluids on a plot of either pressure and volume or pressure and temperature. For an example of the behavior of a pure hydrocarbon substance as temperature and pressure are varied, let's imagine that we fill a pressured container with ethane at 60 °F (289 K) and 1000 psia (6895 kPa). Under these conditions, ethane is in the liquid state. If mercury can be pumped into or released from our container (or cell), we can vary the volume available for the ethane. Figure 1 (Pressure volume changes in a PVT cell) shows how the pressure in the cell falls as the volume is increased, until the first bubble of gas appears. This is called the bubble point. Further increase in cell volume will be at a constant pressure. At any point along this constant pressure line, the amounts of ethane liquid and ethane gas are In equilibrium. The volume of gas in the cell increases at this constant pressure until the point is reached where all the liquid is vaporized. This is called the dewpoint. As the cell volume is increased further at constant temperature, the ethane gas expands and the pressure drops even more.

Figure 1 shows the difference in the slope of the line above the bubble point (very steep) and below the dew-point (more gradual). A liquid is much less compressible than a gas for a given change in pressure. Figure 2 (PVT diagram for a single- component system) shows how the temperature axis can be added to our graph to describe the phase behavior using three variables. The locus of bubble points obtained at various temperatures projected onto the pressure-temperature plane is a line called the vapor pressure curve. When the pressure and temperature are above this line, the ethane exists in the liquid phase; beneath this line the ethane exists as a gas. The vapor pressure curve for a single-component system such as this one ends at the critical point. For ethane, the critical point is about 90 °F (300 K) and 710 psia (4880 kPa). As the critical point is approached, the properties of the gas and liquid phases become similar, and they become identical at the critical point. For a single pure component, the bubble point and dewpoint at each particular temperature correspond to the same pressure. Between the bubble point and dewpoint, however, the relative amounts of liquid and gas coexisting in equilibrium will depend on the specific volume of the gas and liquid. In Figure 3 for example (Pressure-volume plot for pure ethane (from Brown et al., 1948, reprinted by permission of NGAA)), at the 80 °F bubble point, the specific volume of ethane is about 0.0516 ft3/lb (.0032 m3/kg). The inverse of this value is, of course, the density of ethane at bubble-point conditions, about 19.38 lb/ft3 (310 kg/m3). At the dewpoint the specific volume is about 0.138 ft3/lb (.0086 m3/kg) for a density of 7.25 lb/ft3 (116 kg/m3). In between these two points, on the two-phase portion of the line in the pressure-volume plot, the specific volume of the ethane varies between that of the liquid and that of the gas, depending on how much of each is present. The average density of the two-phase material changes, but the density of the liquid and gas does not vary for a given temperature, only the relative amounts of each phase. One can easily see, however, how the densities of the liquid and gas phases become identical as conditions approach the critical point.

Binary System

Reservoir fluids are not pure hydrocarbons. We can begin to understand the added complexity of a multicomponent system by observing a binary system — for example, ethane and normal heptane ( Figure 1 , Pressure-temperature diagram for mixtures of ethane and n-heptane (from Brown et al., 1948)). The first thing we notice is that the bubble-point line and dewpoint line of our pressure volume plot no longer coincide to form the vapor pressure curve. The two-phase region is now apparent on the pressure-temperature plot, as a two-phase envelope. Furthermore, for each possible combination of ethane and heptane, a distinct phase diagram exists. As the percentage of n-heptane in the mixture increases, the critical point of the system moves up to a maximum critical pressure and then down and to the right to a maximum critical temperature, that of n-heptane. When the composition is evenly distributed between components, the two-phase region increases in size, whereas when one component predominates, the two-phase envelope shrinks in size and approaches the vapor pressure curve of that component.

Inside the two-phase region of any particular envelope, amounts of liquid and gas phases coexist in equilibrium. Just as in the case of a single-component system, the pressure and temperature conditions below the dewpoint curve dictate that the mixture exists as a gas, and the conditions above the bubble-point curve dictate a liquid state. An important difference, however, is that the compositions of the liquid and gas phases within the two-phase envelope are not constant. The liquid and gas phases present in the two-phase region will contain more or less of each of the two components, depending on the temperature, the pressure, and the overall composition of the system. For example, if we pick the system from Figure 1 with about 10 weight % ethane and 90% n-heptane, we can illustrate different isothermal pressure conditions, as in Figure 2 (Pressure-temperature diagram for ethane-n-heptane mixture). As we decrease the pressure at constant temperature and travel through the two-phase region, the composition of the gas phase and liquid phase changes. The initial gas phase is rich in the lighter component, but gradually acquires more and more of the heavier component as the pressure is dropped and more gas is present. Concurrently, the liquid phase loses its portion of the lighter component as the pressure drops and the amount of liquid decreases. This change in phase composition is, of course, quite different from the single-component case, where the liquid and gas phases are always 100% pure component. As the complexity of the hydrocarbon mixture increases, determining the composition of the liquid and gas phases becomes increasingly more difficult. Knowing the composition can be important, particularly in the case of gas condensate and volatile oil mixtures.

Multi-Component System

Multicomponent phase diagrams are similar to two-component diagrams, in that the relative amounts of light and heavy components will determine the shape and placement of the two-phase envelope on the pressure-temperature diagram.

Let us consider the broad categories of dry gas, wet gas, condensate, volatile oil, and black oil. The phase diagram for a dry gas, for example, will be similar to that shown in Figure 1 (Phase diagram for a dry gas (after Amyx, Bass and Whiting, 1960; courtesy of McGraw-Hill)). If this hydrocarbon mixture exists in a reservoir with pressure and temperature conditions at point A, the mixture exists as a gas. Because reservoir temperature is usually constant, we can represent the change in reservoir conditions with decreasing pressure as a vertical straight line. This indicates pressure depletion with production. We see that the reservoir fluid remains as a gas, even at lower pressures. If we plot the separator conditions at point B, we observe that even at these conditions the mixture remains a gas — no phase changes have occurred. The kinetic energy of the molecules is high enough so that attractive forces cannot coalesce them into a liquid. This coincides with our definition of a dry gas reservoir as one where no hydrocarbon liquids are present in the reservoir or at the surface.

If we increase the percentage of heavier hydrocarbons in our mixture, the phase envelope shifts and the critical point moves upward and to the right, an effect similar to that of increasing the percentage of the heavy component in our binary mixture ( Figure 2 , Phase diagram for a wet gas (after Amyx, Bass and Whiting, 1960; courtesy of McGraw-Hill)). With a wet gas, the reservoir mixture still exists as a gas; however, production of a portion of the gas to surface separator conditions results in a phase change. By dropping the pressure and temperature, we cross the dewpoint line for our mixture and enter the two-phase region. This accounts for the recovery of condensate liquid in our separator. The precise amount of liquid recovered with a given volume of gas production will depend on the separator conditions and the position of the lines of constant liquid volume inside the two-phase region. Their position depends on the composition of the mixture.

In the case of a dry gas or a wet gas, the reservoir temperature is higher than the maximum temperature of the two-phase envelope. This temperature is called the cricondentherm; it is the highest temperature at which a liquid and gas can coexist in equilibrium. What will happen when the reservoir temperature is lower than the cricondentherm? We can illustrate this situation with an example of the phase diagram for a condensate, specifically a retrograde condensate system ( Figure 3 , Phase diagram showing retrograde condensation (after Amyx, Bass and Whiting, 1960; courtesy of McGraw-Hill)). In this situation the amount of heavy hydrocarbon components in the mixture has increased, pushing the phase envelope to the right. The reservoir temperature is between the critical temperature and the cricondentherm. As the reservoir pressure drops with production, the dewpoint line is crossed and the two-phase region is entered. Normally one expects vaporization to occur with decreasing pressure. In this case, however, the phase envelope bulges out, with a cricondentherm substantially higher than the critical temperature. This allows condensation to occur under conditions that would usually result in vaporization, thus the term retrograde condensation. This phenomenon results from the fact that the attraction between light and heavy hydrocarbon components decreases as the pressure drops and the lighter molecules become separated. Attraction between the heavier molecules becomes dominant, causing them to coalesce and form a liquid. Retrograde condensation continues until a low enough pressure is reached so that the heavy hydrocarbon molecules begin to vaporize and leave the liquid, at which point normal vaporization begins and continues until the liquid once again becomes a gas.

Unfortunately, in the reservoir the production process inhibits the revaporization of condensed heavy hydrocarbons. As condensation begins, gas production continues... indeed, if is the withdrawal of volumes of gas that causes the pressure to drop in our reservoir. Continued gas production, however, causes the overall composition of the reservoir fluid to change. By selectively leaving heavy hydrocarbons in the reservoir, while simultaneously withdrawing the lighter hydrocarbons as gas, we are causing the composition of the remaining reservoir fluid to shift toward a higher percentage of heavy hydrocarbons. This, in turn, causes the phase envelope to shift downward and to the right, requiring increasingly lower pressures for revaporization. The reservoir rock structure complicates matters by requiring that the liquid saturation reach a critical percentage of the pore space before liquid flow can begin. Since this percentage is usually between 10% and 20%, the retrograde liquid often becomes trapped as an immobile liquid phase within the pore spaces of the reservoir. If the reservoir-wide pressure drop with production is kept to a minimum, retrograde condensation may only be a problem in the near-wellbore area, where pressure draw-down is significant. Here it can still cause problems by inhibiting gas flow and reducing productivity. If retrograde condensation occurs on a reservoir-wide scale, significant amounts of valuable hydrocarbon liquids can be left unrecovered in the reservoir.

If our original hydrocarbon mixture has enough heavy components to move the critical point of the two-phase envelope to the right of reservoir conditions, we have a mixture that exists as liquid in the reservoir. Figure 4 (Phase diagram for a high-shrinkage oil (after Amyx, Bass and Whiting, 1960; courtesy of McGraw-Hill)) shows an example of the two-phase envelope for a volatile or high-shrinkage oil. This terminology simply means that the oil contains relatively large amounts of lighter and intermediate hydrocarbons that vaporize easily. Such a composition is reflected in the spacing of the percent liquid lines within the two-phase region. With a fairly small drop in pressure, the relative amount of gas to liquid volume increases rapidly. If our reservoir conditions are such that the mixture exists as a liquid (point A) and a significant drop in pressure is required to reach the bubble point, we have what is called an undersaturated oil. Below the bubble point, a small drop in pressure will cause gas to evolve from the oil. This fluid is then called a saturated oil. The bubble-point pressure is therefore regarded as the saturation pressure for the crude oil system. The gas phase which forms, as pressure drops below this point, is considered as having been dissolved in the liquid phase. Gas produced in this manner is termed solution gas or dissolved gas. As oil is produced from reservoir conditions to surface separator conditions, the amount of evolved solution gas will depend on the amounts of light and intermediate hydrocarbons in the mixture.

Figure 5 (Phase diagram for a low-shrinkage oil (after Amyx, Bass and Whiting, 1960; courtesy of McGraw-Hill )) is a phase diagram for the final common hydrocarbon category, a low shrinkage, or "black" oil. Such crudes are called black oils because of their generally darker, heavier appearance. Low-shrinkage oils require a much greater drop in pressure before a significant amount of gas phase is formed. The percent liquid lines are closely grouped toward the dewpoint line of the phase diagram. This description of a low-shrinkage oil coincides with our earlier statement that there is a relatively small amount of gas phase production at normal separator conditions of pressure and temperature. Numerically, shrinkage may be expressed as (1) a percentage of the final resulting stock-tank oil or (2) as a percentage of the original volume of the liquid. The shrinkage factor is the reciprocal of the formation volume factor, which we will discuss later in this section. Figure 6 (Relationship between shrinkage, shrinkage factor and formation volume factor) gives an example of the relationship.

We can see from our discussion so far that the state of the hydrocarbon system is partially dependent on reservoir conditions. If reservoir conditions are such that our initial pressure-temperature point is within the two-phase region, the reservoir contains an oil accumulation with a gas cap. The gas-cap gas phase is termed associated free gas, and is usually in equilibrium with the oil band it overlies. Such a gas phase is at its dewpoint, and such a liquid phase is at its bubble point. The phase diagrams for each individual phase, and for the total mixture, can be represented as in Figure 7 (Phase diagram for the reservoir fluid mixture and reservoir conditions, which result in a free gas cap and oil accumulation in equilibrium (after Amyx, Bass and Whiting, 1960; courtesy of McGraw-Hill )). A reservoir hydrocarbon mixture, there-fore, may exist as a liquid, a gas, or as an equilibrium mixture of both. Each of these phases has properties that vary with pressure, temperature, and composition.

A.3. Hydrocarbon Gases

Properties of Hydrocarbon Gases

The normal starting point for any discussion of the properties of gases is the ideal gas equation of state. This equation is grounded in several basic laws:

Boyle's law states that at a constant temperature the pressure of an ideal gas is inversely proportional to its volume.

Charles's or Gay-Lussac's law states that at a constant pressure the volume of an ideal gas varies directly with the temperature, and that at a constant volume the pressure varies directly with the temperature.

Avogadro's law states that all ideal gases at a given pressure and temperature have the same number of molecules for a given volume.

The ideal gas equation of state that is derived from these laws is expressed mathematically as:

pV = nRT (2)

where:

p = pressure

V = volume

n = moles of gas

T = absolute temperature

R = constant

The constant R will have different numerical values and dimensions, depending on the type of units used in the equation. However, it is independent of the type of gas and is therefore called the universal gas constant.

This equation is only valid for gases that are "ideal," that is, gases in which:

1. The gas molecules themselves occupy a volume that is insignificant compared to the volume occupied by the gas.

2. There are no molecular forces that cause the molecules to either attract or repel one another.

3. The gas molecules do not lose any energy when they collide with one another.

At very low pressures, many gases exhibit ideal behavior and the equation is valid. At higher pressures, such as those found in oil and gas reservoirs, the equation does not hold. Many attempts have been made to modify the ideal gas law so that it can be used to describe all types of gases over a wide range of temperatures and pressures. The van der Waals equation, the Beattie-Bridgemann equation, and the Benedict-Webb-Rubin equation are examples of equations of state that apply over wider ranges of conditions. However, these equations still lose accuracy at higher pressures and are generally cumbersome to work with (Amyx, Bass, and Whiting 1960).

Real Gas Behavior

The petroleum industry has found that a simple correction factor for the ideal gas equation of state is useful for describing the behavior of hydrocarbon gas mixtures at oilfield conditions. This compressibility factor or z factor, is defined as the ratio of the actual volume of a gas to the volume that would be occupied if the gas behaved "ideally" under the same conditions. This factor varies with pressure, temperature, and composition, and must be determined experimentally. Figure 1 (Typical plot of compressibility factor as a function of pressure and a wide range of temperatures, for a hydrocarbon gas) gives an example of the general variation of compressibility factor, with pressure and temperature, for hydrocarbon gases. One can see that the z factor decreases with decreasing temperature, except in he high-pressure range where the trend reverses itself. At lower temperatures, the isothermal lines have distinct minimums, and the factor decreases with increasing pressure, before starting to increase. At very low pressures, the compressibility factor is close to unity, meaning of course, ideal gas behavior. Adding this correction factor to our equation gives:

pV = znRT (3)

where z = dimensionless compressibility factor

The similarity of the compressibility factor curves for different gases having similar molecular structures led to the development of the law of corresponding states. This law expresses the fact that gases have the same compressibility factor at the same conditions of reduced pressure and reduced temperature. Reduced pressure is simply the ratio of the pressure and the critical pressure of the gas, while reduced temperature is the ratio of temperature and critical temperature. With this law in mind, a generalized chart of compressibility factor as a function of reduced pressure and temperature can be generated and used for any gas. This technique avoids the need for a complex equation of state, but requires that the critical temperature and pressure be known. The critical constants of pure hydrocarbons are published in chemistry handbooks, but determining the critical pressure and temperature for a gas mixture is a more difficult task. To determine the reduced pressure and temperature of a gas mixture, we must rely on another basic law:

Amagat's law states that the total volume occupied by a gas mixture at certain conditions of temperature and pressure is equal to the sum of the volumes the individual pure components would occupy at the same conditions.

It follows from this law that the volume fraction of a pure gas in a mixture is equal to the mole fraction of that gas in the mixture.

We can use the mole fraction composition of a gas mixture to determine the pseudocritical pressure of the mixture by multiplying the mole fraction of each pure component by the critical pressure of that component and adding them together. The pseudocritical temperature of the mixture is obtained in a similar manner. Using these values, we can determine the pseudoreduced pressure and temperature and, thus, the compressibility factor for any given conditions of pressure and temperature. The pseudocritical properties of gas mixtures are used in the same manner as actual critical properties of pure gases when deter-mining compressibility factors. However, pseudocritical constants are not the actual critical constants of the gas mixture. Figure 2 (Compressibility factor as a function of pseudoreduced pressure and pseudoreduced temperature (from Engineering Data Book, 9th ed., 1972; 5th rev. 1981; courtesy NGPSA )is the chart that has been developed to relate compressibility factor to pseudoreduced pressure and pseudoreduced temperature.

Very often, a compositional analysis giving the mole fraction of each component in a gas mixture is not available. In this case, we may use an empirical correlation between gas gravity and pseudocritical properties. Figure 3 (Pseudocritical constants as functions of gas gravity) gives a commonly used correlation that requires only the gas gravity as measured in the field or laboratory.

The gas gravity can also be calculated from a compositional analysis by using Amagat's law to determine the apparent molecular weight of the mixture. Once again, we multiply the mole fraction of each component by the molecular weight of the component, and sum them. This value, divided by the molecular weight of air (28.9), is the gas gravity. However, when a compositional analysis is available, it should be used to determine the pseudocritical constants directly. The correlations in Figure 3 are less accurate and should be used only when a simple gas gravity measurement is the only available data.

The compressibility factor chart in Figure 2 is only accurate for mixtures of hydrocarbon gases. If a natural gas accumulation has significant amounts of nonhydrocarbon components, the compressibility factor calculated using reduced pressure and temperature will be in error. The effect of nitrogen on the compressibility factor has been recorded by several researchers. Eilerts and his coworkers (1948) have presented a method for correcting the z factor for nitrogen concentrations. This method corrects the z factor for an error of less than 1 % with nitrogen concentrations up to 10 mole %, and for an error of more than 3% with nitrogen concentrations of 20 mole % or more. Carbon dioxide has an even greater effect on the compressibility factor. It has been found that 4 mole % carbon dioxide in a natural gas will result in an error of about 5% in the compressibility factor calculated from Figure 2 . Another impurity is hydrogen sulfide. If it is present in small amounts, the critical constants can be used in the usual manner to help determine the pseudocritical properties of the mixture. If large quantities of hydrogen sulfide are present, experimentally determined compressibility factors will be required (Amyx, Bass, and Whiting 1960).

The petroleum engineer is interested in using the gas equation of state primarily for calculating the volume that a quantity of gas will occupy at various conditions of temperature and pressure. The reservoir engineer is particularly interested in relating a volume of gas in the reservoir to the volume it occupies at the surface. Because we are often interested in determining this relationship for a variety of reservoir conditions, it is convenient to fix the surface conditions at some standard reference. Usually, these standard conditions are 14.7 psia (101 kPa) and 60 °F (289 K). However, official standards can be slightly different in different parts of the world or even in different states within the United States. It is important to precisely define the standard conditions of reference, especially when large volumes of gas are being considered.

but, and

where:

z = compressibility factor of gas at reservoir pressure and temperature

T = reservoir temperature

p = reservoir pressure

n = moles of gas

Z_sc = compressibility factor of gas at standard conditions = 1.0

p_sc = 14.7 psia (101 kPa)

T_sc = 60 °F = 520 °R (289 K)

R = universal gas constant 10.732 (8.3143)

thus,

and so,

or,

also

Table 1. Description of gas formation volume factor

Gas Formation Volume Factor

The factor used to relate reservoir gas volumes to standard conditions is termed the gas formation volume factor, Bg. It is described in Table 1, below.

but, and

where:

z = compressibility factor of gas at reservoir pressure and temperature

T = reservoir temperature

p = reservoir pressure

n = moles of gas

Z_sc = compressibility factor of gas at standard conditions = 1.0

p_sc = 14.7 psia (101 kPa)

T_sc = 60 °F = 520 °R (289 K)

R = universal gas constant 10.732 (8.3143)

thus,

and so,

or,

also

Table 1. Description of gas formation volume factor.

This relationship can be calculated by determining the compressibility factor at the desired conditions, using experimental laboratory data, a gas analysis, or a gas gravity. Easy-to-use charts are also available for reading Bg directly for a given gas gravity. These are shown in Figure 1 (Gas formation volume factor for gas gravity of 0.6) Figure 2 (Gas formation volume factor for gas gravity of 0.7), and Figure 3 (Gas formation volume factor for gas gravity of 0.8).

While these correlations are convenient for "quick-look" estimates, they are not as accurate as a calculation using compositional data or gas gravity.

These methods for calculating a gas formation volume factor are only applicable when the reservoir and surface gas have the same composition. If a mixture's phase behavior is such that significant volumes of liquid condensate are produced in going from reservoir to surface conditions, the composition of the gas phase is different in the reservoir and at the surface. Another approach must be taken to mathematically combine the surface fluid volumes and relate them to a reservoir volume.

Gas Compressibility

The formation volume factor illustrates the compressibility of a gas. In Figure 1 (Gas formation volume factor for gas gravity of 0.8) we can easily see that the number of standard surface volumes compressed into a reservoir volume is much greater at high pressures than at low pressures.

However, we can also see that the B g factor changes more rapidly with pressure at lower pressures than at higher pressures. In some engineering calculations we are concerned with this change in volume with pressure at a constant temperature, or isothermal compressibility We should not confuse this property with the compressibility factor mentioned earlier or with the formation volume factor.

The coefficient of isothermal compressibility of a fluid is defined as the change in volume per unit volume per unit change in pressure:

(4)

Because all fluids expand when their confining pressure is released, dV/dp is negative, and the minus sign is added to make the value of compressibility positive.

In the case of gas compressibility, we can combine this definition with the equation of state, to obtain an equation for the gas coefficient of compressibility in terms of pressure ( Table 1 ). At low pressures dz/dp is negative and at higher pressures it is positive. Calculations of Cg for a gas mixture at any given pressure would require that we determine the slope of the z factor isotherms. Trube (1957) has done just that and presented pseudoreduced compressibility as a function of pseudoreduced temperature and pseudoreduced pressure. Figure 2 (Pseudoreduced compressibility for natural gases as a function of pseudoreduced pressure and pseudoreduced temperature (low pressure range)) and Figure 3 (Pseudoreduced compressibility for natural gases as a function of pseudoreduced pressure and pseudoreduced temperature (high pressure range)) show the results. The pseudoreduced compressibility obtained in this manner must be divided by the pseudocritical pressure to yield the gas compressibility, C_g.

Values of isothermal compressibility for a gas are on the order of several hundred times 10-6 psi-1, or in SI units several hundred times 10-7 kPa-1. For example, for a typical natural gas (.66 gravity at 120 °F (322 K)):

at p = 1500 psia (10340 kPa)

c_g = 765 10-6 psia-1 (1110 10-7 kPa-1)

at p = 5000 psia (34470 kPa)

c_g = 94 10 psia-1 (136 10-7 kPa-1)

Gas Density

We can also use the gas equation of state to calculate the density of a gas under any conditions. Table 1, below, shows how this is done. We can see that at a constant temperature, the density of a gas increases with pressure. However, the compressibility factor influences the relationship, and this factor also changes with pressure. The density of a gas is the most commonly measured property. It)'s obtained by experimentally measuring the specific gravity of the gas. Because the specific gravity is equal to the density of the gas relative to that of air, the gas density may be easily calculated.

Beginning with pV = znRT we substitute m/M for n

where:

m = weight of the quantity of gas

M = molecular weight

Thus, pV= zRT

and since density is defined as the weight per unit volume

Table 1. Gas density equation.

Gas Viscosity

Another property of hydrocarbon gases that is often of interest to petroleum engineers is viscosity Any problem dealing with gas flow requires some idea of gas viscosity. Viscosity is a measure of a fluid's internal resistance to flow and is often thought of as a sort of internal friction. The effects of temperature, pressure, and molecular size on gas viscosity can be explained in terms of molecular kinetic energy. As temperature is increased, the kinetic energy of the gas molecules increases, resulting in more frequent molecular collisions and a greater internal friction. This increase in gas viscosity with temperature is the direct opposite of the decrease in viscosity we observe with liquids as the temperature is increased. If the temperature is held constant, an increase in gas pressure causes the gas molecules to move closer together, again increasing the number of collisions and thus the viscosity. The size of the molecules also affects the viscosity. At a given temperature, heavier molecules (high gas gravity) have lower velocities and, therefore, are responsible for fewer molecular collisions. However, as pressure is increased, the molecules are confined together and the heavier molecules have greater attractive forces. This contributes to the number of collisions and thus to an increase in viscosity. The relationship of gas viscosity with temperature, pressure, and gas gravity is depicted in Figure 1 (Viscosity of natural gases containing less than 5% nitrogen, based on methane-propane mixtures for temperatures of and ) and Figure 2 (Viscosity of natural gases containing less than 5% nitrogen, based on methane-propane mixtures for temperatures of and )

Viscosity is defined as the force per unit area required to maintain a unit velocity gradient in a fluid:

(5)

where:

= viscosity

F = force

A = area

v = velocity

y = distance

If a force of one dyne, acting on one square centimeter, maintains a fluid velocity of 1 cm per sec over 1 cm of distance, the fluid has a viscosity of 1 poise. In most cases, poise is too large a unit and centipoise is preferred. One centipoise is 0.001 pascal-second (Pa·s) in SI. Natural gases exhibit viscosities of 0.01 to 0.05 centipoise under typical oilfield conditions.

Measuring gas viscosities at reservoir temperatures and pressures is a difficult procedure. Correlations based on gas gravity are commonly used in place of actual laboratory measurements.

Exercise 1.

You have sampled a newly discovered gas reservoir and are planning its development. Given only the following information, determine the gas formation volume factor and the gas compressibility at original conditions.

Given:

Gas gravity at standard conditions = 0.725
Original reservoir pressure = 5250 psia (36,199 kPa)
Estimated abandonment pressure = 1200 psia (8,274 kPa)
Reservoir temperature = 200 °F (367K)

Determine the incremental recovery possible by dropping the abandonment pressure an additional 500 psia (3448 kPa).

Determine the compressibility of the gas at this abandonment pressure. Where does a unit drop in reservoir pressure result in a larger volume of produced gas — at high pressures or at low pressures?

Solution 1:

Since a gas composition is not available, we must use Figure 1 to obtain values for pseudocritical pressure and temperature. Using the curve for "miscellaneous gases" and our gas gravity we find:

p_pc = 667 psia T_PC = 397 °R

We determine that at original reservoir conditions:

p_pr = p/p_pc = 5250/667 = 7.87

T_pr =T/T_pc =(200+ 460) 1397 = 1.66

From Figure 2 we determine that the compressibility factor (z) is equal to:

z = 1.005 at p = 5250 psia (36199 kPa)

From the relationship B_gi= .00504 (zT/p), we calculate:

B_gi=0.00504 = 0.00064 res. bbl/scf

B_gi= 0.3495 = 0.00356 m³/sm³

To determine the gas compressibility we can use Trube's correlation, given in Figure 3

c_pr = 0.0725 cg = psia-1

We can determine the gas formation volume factor at each possible abandonment pressure.

For p_abn = 1200 psia (8274 kPa) we find:

p_pr = p/p_pc = 1200/667 = 1.8

T_pr = 1.66

z = 0.890 ( from Figure 2 )

B_ga1= 0.00504 = 0.00247 res. bbl/scf

B_ga1 = 0.3495 = 0.01379 m³/sm³

And for p_abn = 700 psia (4827 kPa) we find:

p_pr = p/p_pc = 700/667 = 1.05

T_pr = 1.66

z = 0.930 ( from Figure 2 )

Bga2 = 0.00504 = 0.0044 res. bbl/scf

Bga2 = 0.3495 = 0.0247 m³/sm³

Therefore, recovery may be calculated as:

% recovery =

Thus,

For p_abn = 1200 psia (8274 kPa) RF = 74%

and

For p_abn = 700 psia (4827 kPa) RF= 86%

The incremental recovery obtainable by reducing the abandonment pressure from 1200 psia to 700 psia amounts to about 12% of the original gas in place.

The compressibility of the gas at 700 psia (4827 kPa) is obtained from Figure 4 :

T_PR = 1.66 P_PR = 1.05 c_pr 1.0

cg = = 1500 10-6 psia

Thus a unit drop in pressure at a reservoir pressure of 700 psia (4827 kPa) results in an expansion (production) of 1500 10^-6 volume per volume, almost 14 times as much as at the original reservoir pressure of 5250 psia (36,199 kPa).

For example, if we take 100 ft³ of gas in the reservoir at original pressure and drop the pressure by 10 psia, it results in production of:

100 ft³10 psia 10910-6 ft³/ft³/psia = 30.3 scf

While the samepressure drop at 700 psia results in:

100 ft³10 psia 150010-6 ft³/ft³/psia = 60.7 scf

This behavior helps to explain the attractiveness of adding com pressors to reduce surface pressures and ultimately reservoir pressures in volumetric gas fields.

A.4. Hydrocarbon Liquids

Liquid Compressibility

Liquids are much less compressible than gases. This is evident from the slope of the isotherm on the pressure-volume phase diagram. It takes a much larger increase in pressure to reduce the volume of the liquid phase than to reduce the volume of the gas phase by a similar amount. We can also see that the greater the pressure, the smaller the effect of changes in pressure on the compressibility of the liquid ( Figure 1 , The equation for the coefficient of isothermal compressibility of the oil).

However, as temperature increases, the effect of pressure variations increases. The coefficient of isothermal compressibility for a liquid is given in Figure 1 . In an undersaturated oil reservoir, the coefficient of isothermal compressibility defines the degree of oil expansion that accompanies a drop in pressure. This expansion is an important source of reservoir energy prior to the evolution of gas at pressures below the bubble point. There are correlations available to estimate values of the coefficient of isothermal compressibility for hydrocarbon mixtures, one of which requires the use of reduced temperature and pressure, just as in the case of gas compressibilities.

Liquid Density

The density, or weight per unit volume, of hydrocarbon mixtures in the liquid state is easily determined at the surface via hydrometers that measure the API gravity of an oil. Determining the density of an oil at reservoir conditions can be done in the laboratory or from liquid analyses at both reservoir and surface conditions. Many liquid mixtures follow the additive volume rule: the weight (in pounds) of a component in one pound-mole of a mixture is equal to the product of the molecular weight and the mole fraction of that component in the mixture. Also, the volume of a component in a mixture is the product of the weight of that component in the mixture and the specific volume of that component at the given conditions. The density is then given as the ratio of the sums of the weights and volumes, as shown in Table 1, below. However, reservoir liquids often contain large quantities of lighter paraffins, and the specific volumes of these hydrocarbons are influenced by the attractive forces of the larger hydrocarbon molecules. As a result, the additive volume method of determining reservoir fluid density must be modified to account for the percentage of methane and ethane in the mixture and the molecular weight of the rest of the components. Knowing the density at reservoir conditions can be useful, for example, in determining a pressure gradient within the oil-saturated portion of the reservoir.

(1)	(2)	(3)	(4)	(5)	(6)
Component	Mole fraction in liquid phase xi	Mole wt.Mi	Relative weight,lb/mole, xiMi(2)(3)	Liquid density,lb/cu ftat 60^oFand 14.65psia	Liquid volume, ft³/mole (4)÷(5)
Methane, C₁	0.0019	16.04	.0305	18.70	0.0016
Ethane, C₂	0.0098	30.07	.2947	23.26	0.0127
Propane, C₃	0.0531	44.09	2.3412	31.64	0.0740
Butanes, C₄	0.0544	58.12	3.1617	35.71*	0.0885
Pentanes, C₅	0.0555	72.15	4.0043	39.08*	0.1025
Hexanes, C₆	0.0570	86.17	4.9203	41.36	0.1190
Heptanes plus, C₇₊	0.7681	263	202.0366	55.28	3.6548
Total			216.7893		4.0531

*Average of iso and normal.

Density = = 53.49 lb/ft³ (856.9 kg/m³)

Table 1. Calculation of liquid density from stock-tank-liquid analysis (from Amyx, Bass, and Whiting, 1960; reprinted by permission of McGraw-Hill)

When calculating the density of a liquid, we also need to consider the fact that at a constant pressure hydrocarbon liquids expand when heated and contract when cooled. The thermal expansion of hydrocarbon liquids must be considered, particularly when correcting the volume and density measurements of stock tank oil to standard conditions. For example, the gravity of a crude oil measured as soon as it came out of the flowline might have been 52° API at 185 °F (358K) and atmospheric pressure, for a density of about 48 lb/ft³ (769 kg/m³). In the stock tank at 60 °F (289K), the oil's density will be 51 lb/ft³ (817 kg/m³), about 42° API. Heavier crudes with lower reservoir temperatures will be affected less by temperature. Standing and Katz (1942) have developed correction curves for hydrocarbon liquid density measurements.

Liquid Viscosity

Viscosity is another important property of hydrocarbon liquids. In contrast to gases, liquid viscosity decreases with increasing temperature. This is attributed to the thermal expansion of liquids, which moves the individual molecules away from each other and decreases internal friction. Figure 1 (Viscosity of paraffin hydrocarbons at atmospheric pressure) shows the viscosity of paraffin hydrocarbons at atmospheric pressure and a variety of temperatures. Liquid viscosities are higher than gas viscosities in general, with most oils being in the range of 0.2 to 30 cp or higher, while gas viscosities are in the range of 0.01 to 0.05 cp under typical oilfield conditions.

Liquid viscosity increases as pressure is increased, just as with a gas, although the magnitude of the change is not nearly as great. A typical oil viscosity may increase about 10% per 1000 psi (6895 kPa), while a gas viscosity may increase 40% per 1000 psi. Liquid viscosities are dependent on the composition of the liquid. The higher the amount of light hydrocarbons contained in the liquid, the lower a viscosity under a given set of conditions.

The relationship between pressure and viscosity is reversed when the pressure falls below the bubble point and gas is released from the oil. Now the viscosity increases with decreasing pressure because of the loss from the liquid of the lighter hydrocarbons with lower viscosities.

The same definition of viscosity holds for liquid as for gases. The kinematic viscosity, however, is defined as the viscosity divided by the density:

(6)

The kinematic viscosity plots as a straight line as a function of temperature on a special chart available from the American Society of Testing Materials (ASTM).

Understanding the behavior of hydrocarbon liquids is important for analyzing the performance of reservoirs, for evaluating enhanced oil recovery procedures (where the viscosity of the oil is an important factor), and for the effective design of surface separation and treating facilities.

A.5. Two Phase System

Oil Formation Volume Factors

After the gas phase forms below the bubble point, a further drop in pressure will result in (1) more hydrocarbons entering the gas phase, and (2) an expansion of the gas phase already formed. Because of the compressibility characteristics of gases, this results in a radically increasing volume of gas phase, with continued pressure decline. The liquid volume, on the other hand, shrinks as components escape into the gas phase. The liquid also expands as pressure is decreased, but the loss of molecules to the gas phase overshadows the slight compressibility of the oil and results in a net reduction in volume. This is shown in Figure 1 (Diagram of volume realtionships in the two-phase region). The pressure change the oil undergoes in moving to the surface is much greater than that which takes place in the reservoir as the result of fluid withdrawal. Typically, larger volumes of gas are released at the surface. The liquid and gas phases are withdrawn separately from the separator (or separators) and significant additional volumes of gas may be released in the stock tank, depending on its pressure and temperature relative to the separator.

A volume of reservoir oil at the prevailing pressure and temperature, divided by the smaller volume it occupies at the surface, gives a value that is always greater than unity. This is the oil formation volume factor, B0. If we include with the reservoir oil its initial complement of "dissolved" gas, we obtain the two-phase, or total formation volume factor, Bt. For example, referring to Figure 1 , the formation volume factor at point B (the bubble point), is equal to the "B" volume of oil divided by the "F" volume it occupies after losing some of its mass to vaporization. At some lower pressure, say p2 the total formation volume factor is equal to the oil volume D, plus the gas volume at p2 TR, divided by the volume that the oil volume D will occupy at stock tank conditions. Above the bubble point B0 = Bt. Below the bubble point B0 decreases toward unity, while Bt increases, owing to the release of gas from solution and the continued expansion of gas already released. Using B0, Bt, and the previously discussed Bg, the engineer can describe the volume changes in the overall system. B0 and Bt can be measured in the laboratory or computed from correlations, and are a primary reason for conducting fluid analyses.

Figure 2 (Typical plot of oil formation volume factor and solution gas-oil ratio vs. pressure for an undersaturated oil at constant temperature of 305F) shows a typical plot of B0 as a function of pressure for a fairly volatile undersaturated oil. Above the bubble point B0 actually increases as pressure declines. This is because the oil is expanding, due to the isothermal compressibility of liquids. Below the bubble point, however, the reduction in oil volume due to vaporization of light hydrocarbons (or "gas coming out of solution") masks any expansion of the liquid volume. Also plotted is the solution gas-oil ratio, Rs, the volume of gas per barrel of oil as a function of pressure. Above the bubble point, all the gas remains "in solution" and the value for Rs is constant. Below the bubble point, the amount of "gas" in the "oil" available for release is constantly decreasing as vaporization continues.

Equilibrium Ratios

In order to calculate the compositions of the gas and liquid phases at any given point within the two-phase region, the concepts of vapor pressure and equilibrium ratios should be understood. We remember the vapor pressure curve on our P-T diagram from our earlier discussion of a pure, single-component system. When a pure substance exists at a pressure (vapor pressure) and temperature corresponding to a point on this curve, it may exist as a liquid (bubble point), a gas (dewpoint) or as a two-phase mixture in equilibrium, depending upon the volume of the system. It has been determined that when a special logarithmic plot of pressure and temperature is used, the vapor pressure curves for hydrocarbons become straight lines that can be extrapolated to converge at a common point ( Figure 1 , Vapor-pressure chart of low molecular weight hydrocarbons). Mixtures of hydrocarbons do not exhibit a true vapor pressure, as noted earlier by the skewed surface of the two-phase region in our P-V-T diagram. The bubble point and dewpoint curves do not coincide. With this idea of vapor pressure in mind, we can introduce Raoult's law, which states that in a two-phase mixture at equilibrium, the partial pressure of a component in the gas phase is equal to the product of the mole fraction of that component in the liquid multiplied by the vapor pressure of the pure component. Dalton's law states that the partial pressure of a component is equal to that component's mole fraction in the gas phase times the pressure of the system. Mathematically we can say:

p_i = x_i p_vi (Raoult's Law)

and

p_i - y_ip (Dalton's Law)

therefore

(7)

at pressure p and temperature T

where:

p_i = partial pressure of ith component

x_i = mole fraction of th component in liquid phase

yi = mole fraction of ith component in gas phase

p_vi = vapor pressure of ith component at temp T

K_i = equilibrium ratio

If we define:

n = total moles in mixture

L = moles of material within liquid phase

V = moles of material within gas phase

z_i = mole fraction of ith component in the mixture

x_i = mole fraction of ith component in the liquid phase

y_i = mole fraction of ith component in the gas phase

m = number of components

K_i = equilibrium ratio as defined in equation 7

then, z_in = x_iL + y_iV and = 1

For 1 mole of mixture, V+ L = 1 and z_i = x_i L + y_i V

Since

y_i = K_ix_i

z_i = x_i (L + K_iV)

Since x_i = y_i/K_i

z_i = y_i (L/K_i + V)

Table 1. Equations for determining equilibrium liquid and gas composition.

If we develop another equation which relates y_i and x_i we can begin to calculate their values. This is shown in Table 1. By assuming values for L or V and using the vapor pressures for each individual component at the given temperature, these equations can be solved by trial and error until a summation equal to 1 gives the correct values for y_i or x_i. At the bubble point, L=1 and V=0, since essentially all the mixture is in the liquid phase except for an infinitesimally small amount of gas. Likewise, at the dewpoint, V= 1 and L=0. Therefore we can determine the composition of the dewpoint and bubble-point systems as in Table 2, below.

Bubble point Dewpoint

z_i x_i z_i y_i

z_i = y_i/K_i z_i = x_iK_i

y_i = z_iK_i x_i = z_i/K_i

Liquid mixture in equilibrium with an Gas mixture in equilibrium with an

infinitesimal amount of gas which has infinitesimal amount of liquid which has

a composition given by : a composition given by :

y_i = z_iK_i x_i = z_i/K_i

Table 2. Equations for determining liquid and gas composition at bubble-point and dewpoint conditions.

Unfortunately, in the case of hydrocarbon mixtures, the equilibrium ratio, K_i, is only equivalent to p_vi/p at relatively low pressures and temperatures. Raoult's and Dalton's laws are based on ideal behavior that assumes no variety in intermolecular forces of attraction and repulsion. Such is certainly not the case for the complex mixtures of hydrocarbon components present in reservoir fluids. Additionally, pure components do not have vapor pressures at temperatures above their critical temperature. Many oilfield temperatures, both surface and subsurface, are well above the critical temperatures of many hydrocarbon components. For these reasons, we must rely on experimentally determined K values that vary according to the pressure, temperature, and composition of a system. An example of such data is shown in Figure 2 (Equilibrium ratio data for a gas-condensate system at ) for a gas condensate system, and in Figure 3 (Equilibrium ratio data for a low shrinkage oil at ) for a high-shrinkage crude oil system. Several things can be observed from these charts:

· At low pressures the slope of each curve is about -1.0, as would be expected if K = Pv/P.

Each curve crosses the K = 1.0 line at a pressure that is close to the vapor pressure of the component.

At higher pressures the curves tend to converge to a value of 1.0.

The convergence pressure is the critical pressure when the temperature of the mixture is its critical temperature.

At temperatures other than the critical, the curves do not extend beyond the bubble-point or dewpoint pressure of the mixture, although the curves can be extrapolated to determine the point of apparent convergence. The apparent convergence pressure is similar to the critical pressure of the mixture, in that it is a relative indication of the mixture's composition.

The data in Figure 2 and Figure 3 were experimentally determined for specific mixtures. Equilibrium ratios can also be determined from universal sets of curves that have been developed as charts for each major hydrocarbon component over a range of convergence pressures. The convergence pressure is used as the parameter that represents the composition of the mixture. Figure 4 (Equilibrium ratio data for propane at various temperatures and a convergence pressure of 3,000 psia) and Figure 5 (Equilibrium ratio data for hexane at various temperatures and a convergence pressure of 5,000 psia) give several examples of these charts as published by the NGPA (Natural Gas Processors Association). The NGPA Engineering Data Book also gives a technique for determining which convergence pressure should be used to determine the set of charts to apply for a specific mixture.

While K values are not part of a typical reservoir fluid analysis, the compositional analysis that is routinely done is used to perform gas-liquid composition calculations. These calculations can be important in the design of surface separation equipment, and in the determination of bubble-point and dewpoint pressures and formation volume factors. Gas-liquid equilibrium calculations can be extremely tedious and complicated. For this reason, the charts in Figure 4 and Figure 5 are often incorporated into computer software or data bases for rapid calculation of gas and liquid compositions at various points within the two-phase region for a mixture.

With the advent of high-speed computers, it has also become feasible to use equations of state to describe the vapor-liquid equilibrium behavior of hydrocarbon systems. Equations such as those proposed by Redlich and Kwong (1949) and Peng and Robinson (1976) are available in computer programs and compositional simulators, and are more accurate than the convergence pressure approach to K-values.

Several correlations have also been developed to determine formation volume factors and bubble-point and dewpoint pressures from produced fluid properties.

Exercise 1.

We know that the oil formation volume factor expresses the decrease in volume a quantity of oil undergoes in moving from reservoir to surface conditions. The units of this factor (B_o) are generally given as reservoir barrel per stock tank barrel (res bbl/STB) or reservoir cubic meter per stock tank cubic meter (m³/m³). The value of B_o decreases as reservoir pressure decreases below the bubble point because there is less and less solution gas available in the oil, and, thus, less difference between reservoir and surface volumes. The total formation volume factor (B_t), on the other hand, expresses the reservoir volume of a stock tank volume along with its original share of dissolved gas. Develop an equation for Bt and, using Figure 1 , estimate values for B_t at pressures of 6300 psia (43,440 kPa), 4000 psia (27,580 kPa) and 1000 psia (6895 kPa). Assume a constant gas gravity of 0.60 and a reservoir temperature of 305 °F (425K).

Solution 1:

Since B_o is given in reservoir volume per stock tank volume, express the original dissolved gas volume in similar units and add it to B_o in order to obtain B_g. The volume of released gas that corresponds to the liquid volume B_o is the difference between the initial solution gas-oil ration (R_si) and the solution gas-oil ratio at the specified pressure (R_s). We must multiply this difference by our gas formation volume factor to determine the reservoir volume for the gas. Thus:

Accordingly, we may find the following values:

(1)	(2)	(3)	(4)	(5)
Pressure psia (kPa)	B_o res. bbl/STB (m³/m³)	R_s scf/STB (sm³/m³)	B_g res. bbl/scf (m³/sm³)	Bt res. bbl/STB (m³/m³)
6300 (43440)	2.25 (2.25)	R_si =1920 (341.9)	0.0039 (0.0219)	2.25 (2.25)
4000 (27580)	1.85 (1.85)	1040 (185.2)	0.0055 (0.039)	6.69 (6.69)
1000 (6895)	1.43 (1.43)	300 (53.4)	0.0205 (0.1151)	34.64 (34.64)

Note:Columns (2) and (3) are from Figure 1 ; Column (4) is from Figure 2 .

A.6. Formation Water

Water Compressibility

The compressibility of water is dependent upon the temperature, pressure, and amount of gas in solution (Dodson 1944). The coefficient of compressibility is defined as for other fluids in Equation 4 and here specifically for water;

(8)

Figure 1 (Compressibility of pure water as a function of pressure and temperature, with correction for gas in solution) gives the compressibility of pure water in vol/vol/psi, along with a correction factor that accounts for gas in solution. Several observations are immediately apparent from this graph:

Fluid	Compressibility range
	bbl/bbl/psi	m³/m³/kPa
Water	2-4 10^-6	3^-6  10^-7
Undersaturated oil	5-100  10^-6	7-145  10^-7
Gas at 1000 psi	900-1300  10^-6	1300-1900  10^-7
Gas at 5000 psi	50-200 10^-6	70-300  10^-7

Table 1. Compressibility of water compared to other reservoir fluids (from Craft and Hawkins, 1959; reprinted by permission of Prentice-Hall)

· Water compressibility is smaller than oil compressibility, and much smaller than gas compressibility (see Table 1, above.)

Compressibility decreases with increasing temperature at low temperatures, and increases with increasing temperature at higher (reservoir) temperatures;

Compressibility decreases with increasing pressure;

Compressibility increases with increasing amounts of gas in solution.

The solubility of natural gas in water is itself a function of pressure, temperature, and salinity. Normally this value is about 5 to 25 cubic ft/bbl (1 to 5 m³/m³). Figure 2 (Effect of pressure and temperature on the solubility of natural gas in water, corrected for water salinity) shows the relationship for the solubility of natural gas (.655 gravity) in water with pressure and temperature. Knowing the pressure, temperature, and water salinity allows the determination of gas solubility. Knowing gas solubility, and using the same temperature and pressure, we can determine water compressibility.

Water Formation Volume Factor

The definition of formation volume factor of formation water (Bw) is similar to the oil formation volume factor in that it represents the change in volume a barrel of water undergoes in moving from reservoir to surface conditions. As with oil, the release of gas from solution decreases the water volume, the expansion of the water with reduced pressure increases its volume, and the contraction of the water with reduced temperature decreases its volume. The expansion and contraction due to pressure and temperature reductions are relatively small and offsetting. The solubility of gas in water is considerably less than in oil, generally one hundred times less. The result of this difference is that the curve for water formation volume factor is distinctly different from that for oil formation volume factor, as shown in Figure 1 (Typical graph of formation volume factor for water).

Above the bubble point, the factor increases as the water expands with a reduction in pressure. Below the bubble point, evolution of dissolved gas only partially offsets the expansion trend, so the water formation volume factor continues to increase. The difference between the maximum value of water formation volume factor and the value 1.0 is due to the thermal contraction of the water between reservoir temperature and standard surface temperature. Correlations are available for determining the formation volume factor of water given the pressure and temperature of the reservoir. The correlations assume a .655 gravity gas and pure water. An increase in salinity will produce a decrease in the volume factor, but this change is within 1% throughout a range of formation water salinities from 0 to 300,000 ppm.

Water Viscosity

Formation water viscosity decreases as temperature increases, at a constant pressure. Pressure and salinity changes do not appear to affect wafer viscosity in a regular manner to any great degree (Amyx, Bass, and Whiting 1960). Figure 1 (Viscosity of water at oil field temperatures and pressures) shows the relationship between temperature and viscosity; it can be used to determine viscosity of essentially any salinity water at any pressure.

Formation water properties are required for material balance calculations — particularly when large volumes of water are produced or when large volumes of water-saturated reservoir rock provide energy to the production process. Water viscosity can be an important piece of data when evaluating improved oil-recovery techniques such as waterflooding, streamflooding, and any process that uses water as a driving fluid.

A.7. Fluid Sampling: General considerations

Obtaining Representative Samples

Usually, the main objective of a sampling procedure is to obtain a representative sample of the original reservoir fluid. In designing a sampling procedure, we must consider the effect producing conditions will have had on the reservoir fluids from which we must retrieve our sample.

When the pressure in an oil reservoir drops below the bubble point, gas comes out of solution and forms a separate phase. Similarly, when the pressure in a gas condensate reservoir drops below the dewpoint, liquid begins to condense in the reservoir. In either case, the minor phase must build up to a certain critical saturation within the reservoir rock before it will begin to flow. In the meantime, the composition of the produced fluid is altered by the selective loss of light or heavy hydrocarbons. While the liquids in a gas condensate reservoir may never reach a saturation where they can flow, the gas saturation in an oil reservoir will almost certainly reach a point where gas flow occurs. Because of the relatively low viscosity of gas, this flow of gas will increase rapidly, exhibiting the typical performance trend of a solution gas drive reservoir.

Even if these phenomena are not reservoir-wide, the pressure drawdown associated with flow will often be sufficient to drop the pressure of the fluid in the immediate vicinity of the wellbore below its bubble-point or dewpoint pressure and into the two-phase region ( Figure 1 , Diagram of the pressure distribution within the formation, which illustrates (a) the formation of a free gas saturation near the wellbore when drawdown causes the flowing bottomhole pressure to fall below the bubble-point pressure, and (b) reduced flow rate and reduced drawdown, which insure that a single-phase fluid is flowing that is representative of the original reservoir fluid ).

A sample of such fluid will not be representative of the original reservoir fluid existing farther out in the reservoir and thus will not be suitable for analysis. Steps must be taken to determine the reservoir pressure and temperature, and the general category of the reservoir fluid. If the relationship between reservoir pressure, bottomhole flowing pressure, and bubble-point or dewpoint pressure can be estimated, steps can be taken to ensure that the sampled fluid is representative.

Another concern in obtaining a representative sample is the degree of variation in the original reservoir fluid throughout the reservoir. Large reservoirs having thick vertical oil columns have been known to exhibit variations in fluid properties with depth (Levorsen 1967). Variations such as these cannot be accounted for in a specific sample. A pattern must be established from several samples or producing characteristics, from various wells completed at different intervals. In such cases, proper sampling procedures can ensure that the sample obtained is representative of the reservoir fluid at the sampling depth.

Timing is an important consideration in obtaining a representative sample of the original reservoir fluid. Obviously, it makes sense to sample as early in a reservoir's producing life as possible. Once production creates significant volumes of free gas on a reservoir-wide basis, obtaining a sample of original fluid may be impossible. Often, a reservoir fluid sample will be part of a testing procedure that immediately follows the completion of the first well in a reservoir, for example, in a newly discovered field where development plans may rely on the early determination of expected reserves and production rates. In such cases, it is important that the new well be cleaned up before sampling to remove all traces of drilling fluid from the well and wellbore area. Considerable thought and planning are often needed to coordinate fluid sampling with other testing procedures so that one does not adversely influence the other. For instance, because in modern offshore development situations there is often an emphasis on accelerating production; drilling, completion, production, and testing activities may be going on simultaneously from a single platform. This may certainly affect the time, space, or money allocated to fluid sampling. On the other hand, accurate fluid samples are necessary for the decision-making behind these development activities.

Estimates of fluid properties can be helpful. For example, correlations of bubble-point pressure can be employed with early test data (perhaps from a formation test or drill-stem test) to determine if a reservoir fluid is undersaturated. If it is, a well in that reservoir might be produced for some time without fear of the free gas phase forming and sampling of such a well could be deferred while more critical testing is done.

Product Conditions and Equipment

The producing conditions and surface or subsurface equipment can be important considerations in designing a sampling procedure. The most important of these are:

the type of fluid being sampled;

the stability and accuracy of gas rate, oil rate, and GOR measurements;

the proximity of gas-oil or oil-water contacts to the productive interval;

whether the well is a flowing or pumping well;

the dimensions of downhole equipment;

the well location.

Dry gas reservoirs and highly under-saturated oil reservoirs where the produced fluids remain in a single phase under any flowing conditions (including surface conditions) are relatively easy to sample at the wellhead. An oil reservoir at or slightly above the bubble point will undoubtedly yield free gas at bottom-hole flowing pressures and require conditioning prior to sampling. Conditioning is a procedure whereby the production rate is gradually reduced, resulting in successively higher flowing bottomhole pressures. This simultaneously removes the altered fluid from near the wellbore and moves fresh, unaltered reservoir fluid into the area.

If samples of oil and gas are taken at the surface, it is vital that the producing rates and gas-oil ratio be accurately determined in order that the fluids may be recombined in the correct ratios to formulate a representative sample. If the well is not producing with stable GORs, or if separation facilities are not adequate for accurate measurements, a surface recombination sample should not be considered.

Water production can be troublesome, even in small amounts. If possible, no well producing water should be considered for obtaining a representative hydrocarbon sample.

If necessary, a water-producing well may be sampled if precautions are taken to obtain the sample from above the oil-water contact in the well or separator. Wells where gas Coning into the productive interval is (or may be) a problem should be avoided as candidates for sampling.

Flowing wells are the best candidates for fluid sampling. Production rates are more easily controlled on flowing wells, and measuring the bottomhole pressure is impractical on a pumping well. Subsurface sampling on a pumping well requires the removal of the pump and rods. For obvious reasons, wells on continuous gas lift are unsuitable for surface sampling procedures. However, if a gas-lift well will flow at low rates on its own, it may be conditioned and sampled the same as any flowing well.

The wireline subsurface sampling tool is not extremely large, but may be unsuitable in wells with tubing restrictions (subsurface safety valves, down-hole chokes, and the like), or twisted tubing. Any completion equipment that prohibits the sampler from reaching the producing interval will complicate the subsurface sampling procedure.

Wells located on offshore caissons without deck space or wireline equipment will require special equipment for subsurface sampling work. Wells such as these often have their production commingled at common separation facilities, and special care must be taken to ensure that surface separator samples are representative of the desired well.

Well Conditioning

The objective of well conditioning is to replace the non-representative reservoir fluid located around the wellbore by displacing it into and up the wellbore with original reservoir fluid.

A flowing oil well is conditioned by producing it at successively lower rates until the non-representative oil has been produced. If the well to be sampled has been flowing for at least 24 hours, the oil rate should be measured, along with the flowing bottomhole pressure and GOR.

The production rate is then reduced (by perhaps 30% to 50%) and the gas-oil ratio measured until it stabilizes. This procedure is repeated until a trend in gas-oil ratio is established. The GOR may remain constant, decrease, or increase. If the GOR remains constant, the flow into the wellbore is single-phase, unsaturated oil and the well can be considered ready for sampling. If the GOR decreases, the presence of a free gas saturation is indicated. This gas may be present due to coning (the drawing down of the free gas cap into the producing interval) or due to the flowing bottomhole pressure being less than the bubble-point pressure. Correlations can be used to determine the normal gas-oil ratio to be expected without any free gas production. If the GOR increases, simultaneous production of a gas and oil zone may be indicated. The lower drawdown allows less oil and relatively more gas to flow from separate intervals. Such a well should not be sampled, because it is very difficult to determine when it is adequately conditioned.

The well is considered to be conditioned and ready for sampling when further reductions in rate of flow have no effect on the stabilized gas-oil ratio. At low flow rates, some wells will "head," or produce slugs of liquid followed by gas. This irregular flow makes it difficult to measure the GOR accurately. Some wells may have such low productivity that even a low flow rate requires a large drawdown. Reducing the drawdown enough to bring the flowing bottomhole pressure above the bubble-point pressure may result in "heading," or else take an impractically long time. Replacing the tubing with a smaller diameter string can sometimes increase the flow velocity enough to eliminate heading at low rates.

Pumping oil wells are conditioned in the same general manner. If preliminary correlations show the reservoir fluid to be saturated, the pumping rate should be reduced in order to allow the pressure at the formation face to increase. After the GOR stabilizes, the well should be pumped for several days before taking surface samples. If subsurface sampling is to be done, the pump must be stopped after the well is conditioned and the rods and pump pulled. The well can then be swabbed at a low rate to ensure a representative sample in the bottom of the well before the bottomhole sampler is lowered to the sampling point.

A gas condensate well is also conditioned by flowing it at successively lower flow rates and monitoring the GOR. The GOR should generally decrease as the rate is decreased. This is because the lower rate results in a lower drawdown, which brings the wellbore pressure back out of the two-phase region. The heavier hydrocarbons will be produced rather than condensed in the reservoir, thus increasing the liquid volume at the surface and decreasing the GOR. When the GOR stabilizes, the well has been conditioned for sampling.

The duration of the conditioning period depends upon the volume of reservoir fluid that has been altered as a result of producing the well below the bubble-point pressure, and how quickly it can be produced at low rates. Most oil wells that have not been produced for a long period of time require little conditioning; however, some wells may require up to a week of conditioning to achieve stable GORs.

During the conditioning process, careful records should include:

flowing bottomhole pressure and temperature (if possible);

flowing tubing pressure and temperature;

oil and gas flow rates;

gas meter run data;

separator pressure and temperature;

stock tank oil production rate;

water production rate.

Any auxiliary data should be noted, such as radical surface temperature changes, equipment malfunctions, and measurement methods. API HP 44 includes a model data form for collecting all of this information.

A.8. HC Sampling

Hydrocarbon Sampling Procedures

After conditioning, reservoir fluid samples may be taken with a subsurface sampling device, or individual samples of oil and gas may be taken at the surface and recombined to obtain a representative sample. The choice of sampling technique is influenced by:

the volume of sample required;

the type of reservoir fluid to be sampled;

the degree of reservoir depletion;

the surface and subsurface equipment.

Subsurface sampling is the trapping of a volume of fluid in a pressurized container suspended on a wireline inside the well to the productive interval. This method is often used when:

only a small volume of fluid is required;

the oil to be sampled is not so viscous that it impairs sampler operation;

the flowing bottomhole pressure is known to be greater than the reservoir oil saturation pressure;

the subsurface equipment will not prevent the sampler from reaching the appropriate depth or make its retrieval difficult.

Surface sampling involves the taking of samples of separator oil and gas, along with accurate measurements of their relative rates, and reconstructing a representative sample in the laboratory. This method is often used when:

a large volume of both oil and gas are required for analysis (as in the case of gas condensate fluids);

the facilities for separating oil and gas and measuring their rates are in excellent condition and operated by thoroughly competent people;

the fluid at the bottom of the well is not representative of the reservoir fluid (i.e., gas condensate reservoirs and oil reservoirs producing large quantities of water).

Generally, a subsurface sample is preferred if gas and oil surface measurement capabilities are in question. However, if they are reliable, the surface sampling technique can give a statistically valid value of GOR measured over a long period of time.

Subsurface Sampling

The specific procedure to follow for subsurface sampling includes the following steps:

1. Condition the well to insure that a single-phase, representative fluid is flowing at the productive interval

2. Either shut in the well or allow it to continue flowing at a very low rate

3. Run pressure and temperature surveys to determine fluid levels and pressures

4. Select the sampling point and run the bottomhole fluid sampler to depth

5. Actuate the sampler and retrieve the sample

6. Repeat the sampling operation to obtain duplicate samples (at least two, preferably three samples should be retrieved)

7. Perform a quality check on the samples at the surface

8. Transfer the samples to a storage container for transport to the laboratory

Let us briefly consider each of these steps. The decision to shut the well in or produce during sampling is dependent on flowing conditions. If the maximum flowing bottomhole pressure is less than or close to the reservoir fluid's bubble point, shutting in the well will allow the pressure to build up in the wellbore; ideally, this will redissolve the gas that has formed near the well. The shut-in time may vary from 2 to 3 hours for a high-productivity well to as much as 72 hours for a low-productivity well (Amyx, Bass, and Whiting 1960). If the flowing bottomhole pressure is well above the bubble point, the low-flow-rate technique provides additional assurance that the fluid in the wellbore is representative of the reservoir fluid.

A pressure-temperature survey is run to determine the location of the gas-oil and water-oil interfaces where the well is shut in. These contacts can be determined by plotting the measured pressure versus depth, and noting the points of slope change ( Figure 1 , Typical wellbore gradient plot showing location of gas-oil and oil-water contacts within wellbore).

The best place to sample is at the lowest point in the wellbore passed by all the fluid entering the well, and where the static pressure is above the estimated bubble-point pressure. When water is present, the sample may be collected just above the oil-water contact, if the pressure at that point is at least equal to the bubble-point pressure. If not, the water level may be lowered by bailing. If the water cannot be removed, or if an obstruction prevents the wireline sampler from reaching the required depth, another well should be considered.

The bottomhole sampler is next run to the sampling point and actuated. A variety of commercially available samplers differ primarily in the following ways:

The actuation mechanism (jar-operated, clock-operated, or electronic)

The size of sample obtained (usually 500 to 1000 cm3, with larger sizes available on request)

The working pressure (usually up to 10,000 psi, with higher pressure devices available on request)

The type of construction (normal or corrosive, H2S service)

Figure 2 (Schematic of a typical bottomhole sampler showing (a) sampler open while going in hole, and (b) sampler closed after reaching sampling point) is a schematic of a typical bottomhole sampling device. It is representative of the two principal types of samplers, the Wofford and the Leutert. The sampler shown is about 6 ft (2 m) in length and about 1 1/2 in (4 cm) in diameter. The actuator, which releases the springs and closes the sampler valves, may be mechanically operated by jarring the tool so as to shear the pins holding the sampler open, or may have a clock-driven mechanism designed to close the sampler after perhaps 2 1/2, 5, or 10 hours. Electronic triggering devices are also available. These devices are run on an electric line and allow subsurface pressure and temperature conditions to be monitored at the surface via the electronic cable. When the pressure conditions are considered appropriate, the sampler can be activated electronically from the surface. If run in conjunction with a casing collar locator, precise control of sampling depth is maintained. This approach can be useful in eliminating the need to run a pressure/temperature survey prior to running conventional sampling equipment.

The sampler should be run with the chamber open, at speeds between 100 to 200 ft/min (30 to 60 m/min). Just before reaching sampling depth, speed should be reduced. With the sample at the proper depth, upward and downward movements should be made over a 20 to 30 ft (6 to 8 m) interval to ensure the sampler ports are clear. If a clock-actuation mechanism is used, the sampler should be at the sampling depth about 30 minutes before closing and removed about 15 minutes later (Flopetrol 1973).

At least two or three samples should be obtained. This allows comparison to determine if the samples are truly representative; it also provides a backup sample if leakage or loss occurs on the way to the laboratory. Usually, two complete sampling devices are available on location in case of malfunctions.

A check on the quality of the sample should be done on each oil sample when the sampler has been retrieved to the surface. This check estimates the bubble-point pressure of the fluid in the sampler, allowing its value to be compared with that estimated by correlations. If the measured and estimated values are close, one can feel relatively sure that the sample is good. Also, if multiple samples have measured bubble-point pressures that are similar (within 2%), the samples can be assumed to be representative. The bubble-point pressure measured at the surface should be slightly less than the bubble-point pressure at the point where the sample was taken, simply because the pressure in the sampler has fallen slightly as the temperature has been reduced in bringing the sampler to the surface. If the measured bubble-point pressure is greater than the sampling pressure, it is an indication that the sampler has either collected free gas or leaked oil.

The check is done by removing the actuator from the sample chamber and attaching a head that allows the injection of mercury into the chamber. A calibrated mercury pump is hooked up to the sampler head and mercury is pumped under pressure, opening the sampler valve and allowing mercury to enter the chamber. Small volumes of mercury are pumped and the chamber is agitated until the pressure observed at the pump stabilizes. The pressure and pump readings are recorded. This is repeated until the response from the mercury injection becomes very large. The pump reading versus pressure is plotted, as shown in Figure 3 (Plot of mercury pump reading versus chamber pressure indicating bubble-point pressure of oil samples). A significant change in slope will occur at the saturation pressure.

The sample is then transferred to a container suitable for transportation, again using the mercury pump. Figure 4 (Apparatus for transferring bottomhole sample into storage container for shipment to laboratory using mercury pump pressure) shows how this apparatus is constructed. A container whose capacity should be at least 10% greater than the sampler chamber is filled with mercury while in the laboratory. At the site, the sampler is connected to the system via the transfer head. With valves 4, 5, and 8 open, mercury is pumped until it appears at a loosened connection at valve 6. This purges the connecting lines and head. The connection at valve 6 is tightened and valves 5, 6, 9 are closed, while 4 and 8 are open. Mercury is pumped until the sampler opens. The pressure is recorded, and mercury is pumped again until the pressure is at least 500 psi (3450 kPa) above the sampling point pressure. The sampler should be gently rocked and pressured with mercury until the pressure response to mercury injection shows that the sampler is liquid full. When this occurs, first valve 5 and then valve 6 are slowly opened while pumping mercury to maintain pressure. Valve 9 is then opened slightly and mercury is allowed to drain into a calibrated receiver while pump pressure is maintained. When the volume of mercury withdrawn from the storage container exceeds the volume of the sampler, valve 9 is closed, followed by valve 6. Container pressure and temperature are recorded. A small amount of mercury should be withdrawn (about 10% of sampler volume) to allow a vapor phase to form in the storage container. (Under no circumstances should any hydrocarbon fluid be withdrawn.) This volume should be recorded, and the container tagged and checked for leaks before shipment.

Transfer of the bottomhole sampler fluid may also be carried out using gravity and the density differences of oil and mercury. Figure 5 (Apparatus for transferring bottomhole sample into storage container, using gravity drainage) shows the apparatus for this method. Air is purged from the lines and sight glass by pumping mercury through them and out through loosened connections. With all connections tightened, valves 6 and 7 are closed and mercury is pumped into the system until the pressure exceeds the pressure in the sampler. Valve 6 is opened and the sample is pressured into a single-liquid phase by alternately injecting mercury and gently rocking the sampler. Valve 7 is opened, allowing mercury to flow by gravity from the container into the sampler and displace the sample into the container. The sampler is tilted to enable the mercury to flow gently downhill without mixing with the oil. At this point, valve 4 may be closed, and the mercury pump removed for other use. The sight glass will be filled with mercury when the transfer is complete, if the container has a larger capacity than the sampler. The valves on the container are then closed and the container detached from the system. A vapor space should be created in the container by releasing a portion of the remaining mercury before the container is tagged for shipment.

The API Recommended Practice No. 44 details additional transfer systems when sampling equipment is slightly different in configuration.

The measurement of the sample bubble-point pressure may be done after transfer to the storage container rather than before, depending on the preference of the sampling company. The vapor space in the storage container allows for thermal expansion without danger of bursting or leakage.

Surface Sampling

The procedure for surface sampling includes the following steps:

1. Condition the well to insure that a single-phase representative fluid is flowing into the wellbore

2. Maintain the final conditioning flow rate

3. Accurately measure and record the GOR

4. Sample the gas and oil streams at the primary or first stage separator and at separator pressure

5. Accurately record sample data and tag for shipment to laboratory

After conditioning the well, the final conditioning flow rate should insure a stable GOR. Generally, the low stabilized rate will not exceed 100 bbl/day (16 m³/day) for an oil well, or 1 MMSCF/day (30 Mm³/day) for a gas well, unless the reservoir has high deliverability. (Moderate deliverability oil wells completed with large diameter tubing can experience heading at low rates, so this must be considered a rough rule of thumb.) It is essential that the gas and oil rates be accurately measured and that the resulting GOR not fluctuate. Every effort should be made to employ experienced personnel for this critical operation.

When multistage separation is used, both the samples should be taken from the highest-pressure separator within one hour of each other. Multiple gas and oil samples should be collected, as with subsurface sampling.

Figure 1 (Schematic of surface production facilities with sampling points noted. Difference in conditions between separator and stock tank require adjustment of GOR) shows a schematic of a basic single separator surface production facility with surface sampling points noted. Because the oil is collected at separator pressure and temperature, and the producing gas-oil ratio often is measured relative to the stock tank barrel, an adjustment may be necessary to determine the actual ratio in which the samples should be recombined. The measured gas-oil ratio must be multiplied by a shrinkage factor from separator to stock tank conditions. This factor is usually determined in the laboratory by "flashing" a small volume of separator oil to stock tank conditions. If the oil rate is measured within the separator, or by a positive volume meter associated with the separator, the GOR adjustment is not necessary.

If multiple stage separation takes place and significant stock tank vapors are recovered, the true GOR will be the sum of the gas volumes collected at the various stages divided by the stock tank liquid volume. If the surface gas and oil samples are taken at the highest pressure separator, the recombination GOR should be determined from the relative amounts of gas and oil at the high pressure separator's conditions.

Surface Gas Sampling

There are three basic methods for sampling the gas stream.

Method 1. Filling an evacuated container.

Method 2. Filling an air-filled container after purging it with separator gas.

Method 3. Filling a liquid-filled container by displacement with separator gas.

For a GOR of less than 1500 scf/bbl, two cylinders of 20-liter capacity should be obtained. Between 1500 and 3000 scf/bbl, three cylinders are required, and above 3000 scf/bbl, four cylinders are necessary.

Method 1 requires that the connecting line between the separator and sample container be purged with gas, and that the gas then be allowed to flow into the container until the desired pressure is reached. Although this method requires that an evacuated container be available on location, it is the recommended method for gas sampling. Pressure on the container should be checked prior to sampling to ensure that a vacuum exists.

Method 2 requires that the container be filled and emptied with separator gas several times in order to purge the container of air. The number of recommended successive purge cycles is inversely proportional to the separator's maximum gas pressure, as shown in Table 1, below.

Method 3 requires that a two-valve sample container be filled with a liquid, preferably mercury or salt water (but glycol or water may be used). The container is kept vertical while the upper valve is connected to the separator and the lower valve is opened to withdraw the liquid. When all the liquid is displaced, the valves are closed and the container is removed for shipment.

Maximum container pressure		Minimum purge cycles
psig	(kPa)
10-14	(70-100)	16
15-19	(100-150)	13
20-29	(150-200)	10
30 - 59	(200-400)	8
60 - 84	(400-600)	5
85 - 149	(600-1000)	4
150 - 450	(1000-3100)	3
450+	(3100+)	2

Table 1.
Recommended gas sample container purging cycles for surface separator sampling (from API RP44. 1966; reprinted by permission of API).

Several problems may be encountered during gas sampling, including:

· entrained oil in the gas stream;

corrosive gases in the gas stream;

condensation of hydrocarbons within the container.

Entrained oil can be eliminated by being certain to sample downstream of a correctly functioning mist extractor. If a mist extractor is not available, a filter should be installed in the line connecting the sample container to the source.

Corrosive gases such as H₂S or CO₂ are easily absorbed by water and will react with steel containers. If such gases are present, the liquid-displacement method should not utilize water, brine, or glycol. Additionally, the gas should be dried by passing it over a suitable drying agent (such as anhydrous calcium sulfate). Nonmetal cylinders should be used if gas is suspected to contain H₂S or CO₂. Recommended sizes for gas-dehydration tubes are listed in API HP 44.

Measurement of H₂S and CO₂ content should be performed at the well site if significant amounts of these components are present. Absorption/adsorption of these components during transportation and storage prior to laboratory analysis can render an otherwise "sour" sample "sweet."

Condensation of hydrocarbons within the sample container can be a problem, especially when the separator temperature is significantly higher than the sample container. This problem can result in a non-representative sample being obtained, especially when Methods 2 or 3 are used for sampling. Use of an evacuated container (Method 1) will not eliminate the condensation, but will make its effect less serious. Other steps include warming the sampling container and maintaining it in a vertical position during purging to allow condensation to drain out.

Surface Liquid Sampling

There are three basic methods for liquid sampling:

Method 1. Displacement of a liquid-filled sample container (mercury, brine, glycol, and water are suitable).

Method 2. Displacement of a gas-filled sample container (separator gas is used).

Method 3. Purging the container with separator oil.

All of these methods have the objective of obtaining a sample of separator oil with no loss of dissolved gas and with no contamination by extraneous gases or liquids. All connections used to conduct fluid from the separator to the sample container must be purged with separator oil to avoid contamination.

Method 1 requires that a sample container, with a capacity of about 600 cm3, be filled with a liquid (mercury, brine, glycol, or water) and connected to the separator as shown in Figure 1 (Sampling apparatus for obtaining separator oil sample by the displacement of a liquid from a prefilled sample container). After purging the sampling line, valve 5 is opened to allow the container to come to separator pressure. Then valve 6 is slowly opened to allow the oil to flow into the container and displace the resident liquid with minimum pressure reduction. When the desired sample volume is collected, valve 6 is closed, followed by valve 5. A vapor space is left in the container by withdrawing an additional 10% of the displacement fluid. It is critical that absolutely no amount of hydrocarbon sample be withdrawn during this procedure. The container is disconnected, checked for leaks, and tagged for shipment.

Method 2 requires that the container first be filled with separator gas at separator pressure. Method 2 for obtaining gas samples should be followed for filling the container. With a pressure gauge installed on the top of the container, the bottom of the container is connected through a purged line to the separator. The sample is taken by opening the bottom valve and bleeding off the gas through the top valve, adjusting the rate so that no appreciable pressure drop occurs in the container. A pressure drop will cause part of the oil to vaporize, resulting in a nonrepresentative sample. When oil flows from the top valve, both valves are closed and the container is disconnected. A vapor space may be provided for by displacing a portion (10% of sample container) of the oil with separator gas, at the constant container pressure, from top to bottom.

Another technique is, with the container in a vertical position, to release 10% of the sample contents from the bottom valve. This technique should only be used if the earlier methods cannot be followed. It is very important that the technique used for providing a vapor space be noted on the sampling records. Laboratory personnel must remove the displacing gas before using the oil for recombination.

Method 3 for liquid sampling requires that separator oil be flowed from bottom to top through the sample container at a pressure equal to the separator pressure. After several container volumes have flowed through the container, the valves are closed. A portion (10%) of the container volume is quickly released from the bottom valve with the container in a vertical position. This method requires considerable care to ensure that dissolved gas is not released into the container during flow. Methods 1 and 2 are preferable. The loss of sample in Method 3 can result in questionable results, and this technique should only be followed as a last resort.

Several problems can occur during oil sampling, including these possibilities:

· Separator temperature may be lower than that of the sample container, causing vaporization within the container

H₂S or CO₂ can react with water and attack the container walls

The sample container should be cooled to separator temperature or below. If the oil vaporizes within the container, an insufficient volume of oil may be recovered. Also, if flashed vapors are lost, a nonrepresentative sample will result. H₂S and CO₂ corrosion can be minimized by taking care to obtain a water-free sample from the separator. A calcium sulfate drying tube can be used if emulsified water is unavoidable.

Dry gas wells and oil wells that flow undersaturated liquid at the surface may be sampled at the wellhead using these surface sampling techniques, provided the pressure rating of the sample containers is sufficient to permit such a procedure.

Split-stream sampling is another surface sampling technique that has been utilized, particularly for gas condensate wells. A small-diameter tube is inserted into the middle of the flow stream, usually in the center of the flow stream, 8 to 10 ft (2-3 m) below the wellhead, or just upstream of the separator. The velocity of the fluid in the flow tube should be the same as that in the flow line or tubing. The flow is diverted into sampling bottles or into a small, temperature-control led separator, from which recombination samples may be taken, thereby averaging out any small fluctuations in the instantaneous GOR. The split-stream method loses accuracy with very high liquid content gases where much of the liquid is flowing close to the walls of the tubing, and a sample from the tubing center is not representative (Amyx, Bass, and Whiting 1960). For this reason split-stream sampling is not recommended and its use is discouraged for any sampling involving multiphase flow.

Exercise 1.

Given the following information about these wells, suggest the appropriate sampling procedure.

Well # 1: Flowing well in new reservoir with a one-week production history. Straight hole with no tubing restrictions.

Oil rate: 200 STB/day (31.8 m3/day)

GOR: 300 scf/STB (stable) (53.4 sm3/m3)

Oil gravity: 30 °API (specific gravity = 0.876)

Gas gravity: 0.60

Water cut: 0

Flowing tubing pressure: 800 psia (5520 kPa)

Flowing bottomhole pressure: 2000 psia (13,790 kPa)

Avg. reservoir pressure: 2200 psia (15,170 kPa)

Reservoir temperature: 150 °F (339K)

Bubble-point pressure (est. from correlations): 1800 psia (12,410 kPa)

Well #2: Flowing well with several months of intermittent production history. GOR has been relatively stable, no increasing trend.

Gas rate: 7 mmscf/day (198.2 Msm3/day)

Oil rate: 140 STB/day (22.3 m3/day)

GOR: 50,000 scf/STB or 20 STB/MMscf (8905 sm3/m3)

Oil gravity: 55 °API (specific gravity = 0.759)

Gas gravity: 0.75

Water cut: trace

Solution 1:

Well #1 appears to be producing from an undersaturated reservoir. The flowing bottomhole pressure appears to be above the estimated bubble-point pressure and a representative fluid can be assumed to be flowing into the wellbore from the reservoir in the vicinity of the well. The flow into the wellbore is single-phase unsaturated oil and the well can be considered ready for sampling. Either surface or subsurface sampling can be carried out.

This well would be an ideal candidate for subsurface sampling because:

1. Subsurface sampler volumes are sufficient for black oil studies;

2. Even though the surface production equipment may be new and fairly accurate, a subsurface sample is preferable to a recombined sample;

3. No tubing restrictions are present to prevent wireline activity.

However, the final decision to use a subsurface or surface sampling procedure will be influenced by:

1. The relative cost of surface vs. subsurface procedures;

2. The location of the well and the availability of wireline and surface production equipment;

3. The risk of shutting in the well and running wireline equipment in hole.

Assuming a subsurface sample is decided on, the procedure that we would use is as follows:

1. Shut in well or flow at low rate;

2. Run pressure and temperature surveys to determine fluid levels and pressures (if well is shut in;

3. Select sampling point and run sampler to dept;

4. Actuate sampler and retrieve; repea;

5. Perform quality check on samples at the surface;

6. Transfer samples for transport.

Well #2 is a gas condensate well. Because the GOR does not appear to be increasing, the chances are good that the dewpoint has not been reached and the flowing bottomhole pressure is greater than the dewpoint, meaning that a single-phase fluid (gas) exists in and around the wellbore. This fluid should be representative of the reservoir fluid.

Sage and Olds's correlation is not really adequate for estimating dewpoint pressure in this case. We must rely on the GOR behavior to assure us of a representative sampling, and must measure the dewpoint in the laboratory.

Subsurface sampling is not advisable. The recommended procedure is as follows:

1. Stabilize GOR and flow rate at the lowest possible flow rate (in order to have the minimum drawdown) but compatible with well and separator stability and homogeneous flow in the tubing. The minimum flow rate required to remove liquids from tubing by gas flow may be estimated by the following equations (from Turner, Hubbard, and Dakler 1969):

qg (MMscf/day) =

where:

A = flow area of tubing, ft2

p = well head pressure, psia

T = well head temperature, °R

q_g = gas flow rate, Mmscf/day

z = compressibility factor

v_g = gas velocity for average gas gravity (0.6) and average temperature (120 °F)

If, for example, our well has a flowing tubing pressure of 1500 psia (10,340 kPa), and 120 °F (322 K) with 2 in tubing, the minimum flow rate would be 1.6 MMscf/day (45.4 Msm³/day), assuming a 7 factor of 0.765.

2. Accurately measure and record surface flow rate, GOR, pressure, and temperature data.

3. Sample the gas and oil streams at the primary separator using Method 1 (evacuated container) for gas sampling and Method 1 (displacement of a liquid-filled sample container) for oil sampling. Multiple samples of each fluid should be obtained. At least 1000 cm³ of separator liquid is needed for the PVT analysis.

4. Tag container and prepare for transport.

Exercise 2.

List and describe the problems that can occur during surface sampling of separator fluids. What problems might you expect during a subsurface sampling procedure?

Solution 2:

Surface fluid sampling problems include:

1. Oil in the gas stream caused by inefficient separation. If the separator does not have a correctly functioning mist extractor, a filter may need to be installed in the line connecting the sample container to the separator. If this problem is significant, gas volume measurements may be unreliable;

2. Corrosive gases in the gas stream. H₂S and CO₂ will react with water and steel. The gas should be collected in an evacuated container after being dried over a suitable drying agent. H₂S and CO₂ measurements should be performed on site to prevent the generation of erroneous values;

3. Condensation of hydrocarbons within the gas sample container. Use an evacuated container warmed to separator temperature if possible;

4. Vaporization within the oil-sampling container. Maintain container at separator temperature or lower;

5. H₂S and CO₂ corrosion. Obtain a water-free sample.

Subsurface fluid sampling problems include:

1. Wireline sampler operation problems. Be sure that all tools are tested and in good working order prior to use. Make sure that a complete and accurate well schematic has been used to compare tool OD's and tubing ID's prior to running in the hole. A dummy run may be necessary if some question exists about the tubing's condition;

2. Accidental sampler closure. This can be avoided by running into the hole at reasonable speeds and being careful to avoid abruptly hitting a liquid level if the well is shut in. If a clock mechanism is used, 30 minutes should be allowed for after reaching sampling depth;

3. Sampler leakage. The quality check and bubble-point determination can help detect this. A new sampler may be required or portions of the chamber seals may need to be rebuilt.

Water Sampling

The goal of formation-water sampling is also one of obtaining a representative sample. The composition of formation water is not generally so dependent on the temperature and pressure changes, and sampling procedures are in most cases simpler than for oil and gas. Small volumes of formation water retrieved from a drill-stem test or formation tester can often be used for analysis of dissolved salts. Samples are more often taken from the separator or wellhead.

Sampling Methods

Drillstem tests are temporary completions of a productive interval; they allow short-term production of reservoir fluids for the measurement of pressure and collection of flow-rate data. During a drillstem test, or DST, fluids may or may not flow to the surface, depending on the reservoir pressure, the productivity of the formation, and the design of the test. If formation fluids flow to the surface, sampling may proceed just as in the case of a producing well test, as long as care is taken to produce all of the drilling mud filtrate that has invaded the pore space in the near wellbore area. This cleaning-up process may take some time, depending on how much filtrate loss or lost circulation occurred during drilling. DSTs are not generally reliable means of obtaining oil and gas samples, unless all the requirements for a representative sample are met. Water samples taken from DSTs are generally reliable if tests are made on site to ensure a representative sample. For example, when water samples that had been taken at intervals from the produced fluid column within the drillpipe on a well that did not flow to surface were analyzed, they showed how errors can be caused by incorrect sampling. In Table 1, below, we see an analysis of top, middle, and bottom samples taken from a 50 ft (15 m) column of fluid. These data show an increase in salinity with depth in the drillpipe, indicating that the first water was contaminated by mud filtrate (Noad 1962; Collins 1975). Only the bottom sample is representative.

Constituent	Concentration (mg/liter)
	Top	Middle	Bottom
Sodium	29,600	43,500	71,800
Calcium	8,100	13,100	22,400
Magnesium	600	900	1,400
Bicarbonate	500	500	400
Sulfate	2,000	1,300	500
Chloride	59,900	91,800	154,000
Total dissolved solids	100,700	151,100	250,500

Table 1. Drillstem test recovery of Smackover Limestone formation water. Changes in ionic composition are apparent as produced water becomes more representative of formation fluid in the lower portion of the drillpipe (after Noad, 1962; courtesy of JPT)

If it is desirable to measure the dissolved gas content of the formation water, the drillstem must be shut in to allow the pressure at the bottom of the well to build up. A sample of the formation water at formation conditions may then be taken with a subsurface sampler run through the drillpipe. Some DST tools also contain a subsurface sampling device within the tool assembly that catches a sample at bottomhole flowing conditions.

Sampling at the flow line is conducted in a manner similar to Method 3 for separator liquid. Figure 1 (Apparatus for sampling water at the flowline) illustrates how several volumes (at least ten) of water are allowed to flow upward through the sample container. The sample container valves are then closed. If any air bubbles are present in the water, a new sample should be taken.

Figure 2 (An illustration of the two techniques for obtaining air-free water samples: (a) overfilling sample container from the bottom up and (b) overfilling sample container within a larger container and capping under water) is an illustration of two methods of sampling water from the wellhead. In the first, a plastic or rubber tube is used to fill a sample bottle from the bottom. Several volumes of fluid are displaced before the tube is slowly removed and the container sealed. An alternative method is to place the sample container within a larger container, filling from the bottom of the inner one until the brine overflows both containers. The sample is then capped under water to prevent air contamination.

At wellheads where oil and water are produced together, a simplified surge-tank separator can be assembled, as shown in Figure 3 (Apparatus for obtaining an oil-free sample at the wellhead). The separation container is first rinsed with well fluid and then filled from the bottom. An oil-free sample is then withdrawn at the bottom of the separator.

Sample Containers

Containers that are used for water sampling are made from plastic, rubber, metal, and boro-silicate glass. Glass can absorb iron and manganese and may contribute boron or silica to the sample. Metal containers can yield abnormally high iron content values. Plastic or rubber containers may similarly contribute organic compounds to the water (Collins 1975).

Probably the most satisfactory container, especially if the sample must be stored before use, is the polyethylene bottle. Because metal catalysts are used in plastics manufacturing, the metal content of the polyethylene should be obtained from the manufacturer. High-metal-content plastics are not suitable (Collins 1975).

A pH determination should usually be made at the sampling point. The pH can be affected by the sampling technique and by handling in the laboratory. Changes in pH between source and laboratory can only be exposed by having testing data from the source.

Soluble iron in the water can precipitate out unless care is taken to keep oxygen out of the sample or a solution of hydrochloric acid is added to the sample to "fix" the iron and keep it in solution until analysis (API RP 45 1981). Often two identical samples are obtained in plastic containers one of which is acidified. Because glass containers can absorb heavy metals even if acid is used, plastic is preferable (Collins 1975).

Each sample container should be clearly marked with all pertinent data.

A.10. Sampling: Transportation & Safety

Transportation and Safety

Safety concerns in reservoir fluid sampling procedures can be considerable, particularly when hydrogen sulfide is present or high pressure conditions prevail.

Of primary importance is the need for a vapor space within the liquid sample containers. Thermal expansion of the liquid could cause the container to exceed its pressure limits if the temperature rises. In general sample containers should be kept at reasonable surface temperatures and not allowed to sit in direct sun or placed in hot areas. Care should be taken to protect the container, especially the end valves, during shipping and handling. Government regulations concerning transportation of flammable and pressurized fluids must be followed. The valves on each end of the sample container should be fitted with plugs to prevent accidental opening during transportation.

Corrosive and noxious gases such as CO2 and H2S should be collected in containers constructed of appropriate materials to minimize deterioration of the container. Careful labeling should indicate any H2S content.

Of course, the pressure ratings of all containers, connection tubing, samples, valves, and fittings should be followed.

A.11. Reservoir Fluid Analysis

Flash versus Differential Liberation

A volume of reservoir oil will shrink as pressure is reduced and vaporization occurs. The degree of shrinkage is dependent not only on the temperature and composition of the fluid, but also on the manner in which the separation process is carried out.

Flash or equilibrium separation is the condition that occurs when the fluid's pressure is radically and suddenly changed and the whole system immediately separates into two phases. In the laboratory, this type of liberation of gas is carried out in a mercury cell or in a small-scale separator at surface temperature. It is felt that flash liberation most nearly approximates the situation that occurs in field separators.

Differential liberation is the process in which a gradual decrease in pressure is applied to the fluids with the continual removal of the released gas from contact with the liquid. Such a situation is thought to occur within the reservoir as solution gas begins to flow more rapidly from the reservoir than the oil. In the laboratory, a stepwise procedure of dropping the pressure of a sample volume at reservoir temperature (flashing) and removing the released gas approximates the reservoir process of gradual change. In Figure 1 we see schematics of flash and differential liberation. (Schematic of (a) flash liberation and (b) differential liberation. The degree to which oil volume is effected by the separation process is dependent on the composition of the oil. In the case of a low shrinkage oil (c), differential liberation provides for a larger volume of stock tank oil. A high shrinkage oil (d) is affected differently.) The composition of the reservoir fluid will determine which of the two processes results in a greater degree of oil shrinkage. For most black oils, differential liberation results in less shrinkage.

Multistage separation is an attempt to approach differential separation at the surface in order to achieve a larger volume of oil in the stock tank per barrel produced. There is also an optimum set of separator conditions (pressure, temperature) for maximizing stock-tank-oil volume.

Similar differences exist for flash and differential retrograde condensation of heavy hydrocarbons from a gas. Once again, flash condensation is more representative of surface separation processes, while differential condensation is more representative of reservoir conditions.

PVT Laboratory Procedures and Report: Gas-Crude oil Sample

The laboratory procedure for sample analysis generally follows this procedure:

1. Determine the mole fraction composition of the subsurface sample, or of the individual gas and oil surface samples.

2. Determine shrinkage for surface separator samples.

3. Recombine the surface oil and gas samples according to test data.

4. Perform a relative total volume or PV test.

5. Perform a differential liberation test at reservoir temperature in the test cell.

6. Perform a flash separation test at various separation pressures and temperatures.

7. Determine the fluid viscosity over a range of pressures at reservoir temperature.

The first page or two of a report generally is a cover letter which includes a statement of the field quality check on the samples and of the tests performed. The bubble-point pressure at reservoir temperature, the total solution gas and relative volume factor for differential liberation, the range of oil viscosities, and the optimum separator pressure are usually summarized in the text. Following this is a listing of formation characteristics, well characteristics, and sampling conditions. All of these data are obtained before or during the sampling procedure.

Some testing laboratories may include more extensive amounts of the field test data and conditioning procedures. Most important are the reservoir pressure, gas and oil rates, and reservoir temperature, although all of the data are helpful and possibly critical in interpreting the test results.

Hydrocarbon Compositional Analysts

Next in the report we will find the compositional analyses of the reservoir fluid. If a surface recombination sample has been taken, the composition of the individual separator gas and liquid samples is listed along with the composition of the recombined reservoir fluid. The basis of recombination is always given, as well as the molecular weight and density of the heptanes-plus (C7+) fraction and the properties of separator gas and liquid.

Composition may be determined by fractionation in a low or high temperature fractionating column, by mass spectrometer, or, more commonly, by gas chromatography. A chromatograph ( Figure 1 , Gas chromatograph used for compositional analysis of hydrocarbon mixtures) separates the components according to boiling point, in a special column. Liquids are analyzed by capillary gas chromatography up to a certain molecular weight (C35), and the heavier fraction is characterized by molecular weight and density. Molecular weight measurements are performed on the heavier fraction by cryoscopy. In this technique the freezing point of a mixture of benzene and the oil's heavy ends is measured and compared to the freezing point of pure benzene. This allows a determination of the heavy fraction's molecular weight and density.

As mentioned earlier, the recombination of the gas and oil samples must be made based on the separator GOR expressed with respect to separator liquid. If a separator-stock tank oil shrinkage factor was not measured in the field, it must be determined in the laboratory by recreating separator and stock tank conditions and measuring the difference in volumes of the separator and stock tank liquid. Only on this basis can the samples be recombined and the fluid composition analyzed.

Relative Total Volume Test

After the representative fluid sample is transferred into the test cell, usually a high-pressure, high-temperature, mercury cell ( Figure 2 , High-pressure, high-temperature mercury test cell for relative volume determinations. Viewing mirror allows view glass to be pointed away from operator), the flash separation procedure shown in Figure 3 may be carried out. (Schematic of (a) flash liberation and (b) differential liberation. The degree to which oil volume is effected by the separation process is dependent on the composition of the oil. In the case of a low shrinkage oil (c), differential liberation provides for a larger volume of stock tank oil. A high shrinkage oil (d) is affected differently.) The pressure on the sample is decreased from above reservoir pressure by withdrawing mercury at reservoir temperature. The pressure and volume changes allow the plotting of a graph to determine the bubble point. Once gas begins to appear, the sample is brought to equilibrium after each volume change by rocking the cell. The resulting data are expressed as the relative total volume and are found in the PVT report as a columnar table. The volume is expressed relative to the volume at the bubble-point or saturation pressure. No material is removed from the cell during this procedure.

Usually, the data entered in the PVT report is "smoothed" to account for the errors in reading very small changes in volume. This smoothing is done by fitting the total relative volume data to a "Y" function curve:

(9)

The compressibility of the reservoir oil above the bubble-point pressure is usually presented here also, since it is obtainable from the pressure-relative volume data by calculating the change in volume per unit change in pressure.

Differential Liberation Test

Usually the results of a differential liberation test appear next in our PVT report. Once again, the pressure is incrementally lowered in the test cell, increasing the volume by withdrawal of mercury ( Figure 2 ). When gas has evolved from the oil, it is displaced from the cell at a constant pressure by injecting mercury ( Figure 3 ). The volumes of the free gas and remaining oil are measured at cell conditions, and the free gas is also measured at standard conditions. This procedure is repeated (perhaps 10 or 12 times) until only oil remains in the cell at reservoir temperature and atmospheric pressure. The residual oil is then cooled to standard temperature in order to measure its change in volume. This residual oil is not the same as stock tank oil, because the differential process of gas liberation is not equivalent to flash liberation. Differential liberation is carried out at reservoir temperature, thus the residual oil will have more of the lighter ends "cooked off." This results in the gravity of the residual oil being lower than that of an oil resulting from a flash liberation carried out at separator temperature.

In the differential liberation test, material is removed from the system, changing the overall composition of the hydrocarbon mixture in the cell.

The relative oil volume data presented in the PVT report are given relative to a volume of residual oil at standard conditions. The solution gas-oil ratio is also presented, based on the volume of gas liberated in each incremental pressure drop and the total gas released. The compressibility factor, formation volume factor, and gas gravity are determined for the gas released during each pressure decrement, and presented in the report. The oil density also is reported here.

The relative oil volume data measured so far are not equivalent to the oil formation volume factor we wish to obtain for use in material balance and reserve calculations. In order to adjust the differential liberation data to obtain B0, we need to simulate the flash liberation process that takes place in the separator and results in stock tank oil. This is done in the flash separation test.

Flash Separation Tests

In this test, part of the reservoir fluid is ejected from the test cell at reservoir temperature and saturation pressure into a small-scale stage-separation system. The separation pressures and temperatures are carefully controlled and generally determined by the engineer requesting the test. The volume of gas liberated at each stage and the volume of remaining liquid are measured. Normally the test is carried out for two stages of separation. The first-stage separator pressure is generally varied to include at least four possible separator pressures at ambient temperature, and the second stage is generally carried out under stock tank conditions of o psig (1 atmosphere) at ambient temperature.

The data presented for this test include the GOR in gas volumes per separator barrel and per stock tank barrel. The formation volume factor given here is the volume of saturated oil at bubble-point pressure and reservoir temperature per volume of stock tank oil at standard conditions. The gravity of the oil and the gravity of the flashed gas are also reported for each separator pressure.

These data may be plotted as a function of separator pressure ( Figure 4 , Formation volume factor and oil gravity versus separator pressure) to reveal the optimum separator pressure. The optimum pressure is that which maximizes stock tank oil gravity and volume. Assuming that we will be able to install a separator that operates at this optimum pressure, we can use the flash separation data from the stage separation closest to this optimum pressure to adjust our earlier differential data, thus obtaining B0 for the proposed field conditions. This is done by multiplying each differential liberation volume factor by the ratio of the flash separation volume factor and the differential liberation volume factor at the bubble point. In Figure 5 (Adjustment of differential oil relative volume curve to separator conditions) we see the adjustment. Similarly, the solution GOR curve must be adjusted as shown in Figure 6 (Adjustment of differential solution GOR curve to separator conditions). The Rs and B0 values generated in this manner may be used to predict reservoir performance when surface separation equipment performs as expected. Some laboratories present adjusted formation volume factor and solution GORs in their reports based on the optimal separator pressure. Other labs leave the selection of an optimal separator pressure up to the engineer.

Fluid Viscosity

Reservoir fluid viscosity is measured at reservoir conditions using a rolling-ball viscosimeter ( Figure 7 , Rolling-ball viscosimeter used for determining fluid viscosity at high pressures and temperatures). This instrument electronically measures the time required for a steel ball to roll a given distance through a tube filled with the fluid to be tested. The clearance between the tube and the ball is varied according to the general magnitude of the fluid viscosity. The viscosity of the oil above the bubble point and below the bubble point is measured in this manner. Gas viscosities are generally determined from the gas analysis using correlations. By performing a compositional analysis on the gas liberated at any point during differential liberation or flash separation tests, the gas viscosity can be determined. When measured, gas viscosities are usually measured by determining the pressure drop for flow through a capillary tube.

The PVT report generally includes plots of all the important tabular data, along with a plot of the subsurface pressure survey if a subsurface sample was taken. Although not part of a typical PVT report, the determination of equilibrium ratios for a reservoir fluid may be done at special request. A test cell is charged with a reservoir fluid sample and the sample is flashed by withdrawing mercury and dropping the pressure. Samples of the equilibrium oil and gas are individually withdrawn and analyzed using gas chromatography. The equilibrium ratio at the particular pressure and reservoir temperature can be determined from the analyses. The process is repeated with a fresh sample to a lower pressure than the previous one. This is done several more times over a desired range of pressures.

PVT Laboratory Procedures and Report: Gas Condensate Sample

The laboratory procedure for a gas condensate sample analysis generally follows the oil analysis procedure with a few exceptions:

A gas condensate sample must always be of the surface recombination type because of the collection of liquids in the wellbore;

A compositional analysis of the well stream at various stages of pressure depletion is done;

A calculation of the gas and liquid recovery during pressure depletion is made;

A calculation of the retrograde condensate liquid volume is presented.

The first few pages of the report generally consist of a cover letter that describes the sampling procedure and the tests performed. A table of formation and well characteristics and the sampling conditions then follows. Analyses of the separator gas and liquid samples, along with an analysis of the recombined well stream and the basis for recombination, are next. The reservoir fluid is transferred to a test cell for the remaining tests.

Relative Total Volume Test

This test is performed in a manner similar to that for the gas-crude oil sample, except that a sharp change in the pressure-volume curve is not apparent in a gas condensate system. The dewpoint must be observed in the test cell, as a "fog" or mist forms with pressure reduction. The volume data are presented relative to the volume at the dewpoint pressure. The z factor for the gas above the dewpoint is determined by the change in volume with pressure. Liquid volume data collected during this test can be most helpful if any compositional simulation work is carried out on the subject fluid and reservoir.

Depletion Study

During this study, the test cell is charged with a known volume of reservoir fluid. The pressure in the cell is lowered by producing gas from the top of the cell to simulate the production of the actual reservoir. At predetermined pressures below original reservoir pressure the produced well stream is sampled and an analysis of the fluid is obtained. Compositional data determined during a constant volume depletion study provide equilibrium K-values, which are necessary for any computer simulation of the gas condensate reservoir. Compressibility factors are also determined. With these data, it is possible to calculate the cumulative recovery of condensate liquids per unit volume of original fluid for given surface separation conditions, which are usually two separators and a stock tank. The retrograde condensation within the reservoir may also be calculated and expressed as a percentage of hydrocarbon pore space.

Graphical plots of the produced gas z factor, cumulative volume percent produced, and the amounts of gas and liquids recovered per volume of original fluid are generally included in the report. The PVT report on a gas condensate fluid sample is utilized in the design of surface separation and liquid recovery systems, and in the design of gas cycling programs to recover liquids left in the reservoir.

Volatile oil Analysis

As described previously, a volatile oil is a very light oil with a very high gas-oil ratio. Based strictly on surface production characteristics it may be impossible to determine if the reservoir fluid exists as a volatile liquid or a near-critical gas.

When such a situation is suspected, the fluid sample should be obtained as for a gas condensate reservoir (i.e., surface sampling), and observed in a windowed cell at reservoir temperature to determine whether the reservoir fluid is a volatile oil or a gas condensate. The procedure for analyzing a volatile oil is different than that followed for either a gas-crude oil sample or a gas condensate sample.

Analysis of a volatile oil will typically include:

relative volume test — this pressure — volume relationship is determined as for a "black" oil reservoir fluid;

constant volume depletion study — carried out in a manner similar to that done for a gas condensate sample and allowing the measurement of oil shrinkage and evolved gas composition;

molecular composition of the reservoir fluid — this composition is determined from the chromatographic analysis of the associated gas and the molecular weight of the produced liquid, combined according to the gas-oil ratio;

flash separation and viscosity measurement — determined as for a "black" oil reservoir fluid.

Because volatile oil reservoirs exhibit high shrinkage immediately below the bubble point, a plot of liquid volume as a percent of initial hydrocarbon pore space, as a function of reservoir pressure, is usually determined from the depletion test and included in the PVT report. This is similar to the retrograde liquid curve which is determined for a gas condensate, in that it expresses the relative volumes of gas and liquid phases in the reservoir rock.

The depletion test is necessary to determine the amount of liquid which will be recovered from the rich gas which evolves from the original reservoir oil. Neglecting this condensate will result in an estimate of oil recovery which is overly pessimistic.

Formation Water Analysis

The basic parameters determined in a typical water analysis are:

pH: usually measured with a pH meter immediately after sampling;

resistivity: usually measured directly using electrodes in a resistivity cell or calculated from a compositional analysis;

specific gravity: usually directly measured using a specific gravity balance or hydrometer; and

ionic composition: measured using titration or spectrophotometric methods, depending on the constituent.

The typical ions measured by titration are carbonate, bicarbonate, chloride, and sulphate. Sodium and potassium are generally measured using flame spectrophotometry, in which a small amount of the brine is aspirated into a flame and the light emitted by the excited ions is measured. Calcium and magnesium are measured by titration or atomic absorption spectrophotometry.

Various spectrophotometric and titration techniques are used for determining trace amounts of other ions. Details on these techniques are given in the API Recommended Practice of Analysis of Oil Field Waters RP 45 (1981).

Exercise 1.

List the typical laboratory tests performed on samples of a black oil and gas condensate.

Solution 1:

The laboratory tests performed during a routine PVT analysis are —

For a black oil or gas-crude oil system:

1. Fluid compositions and recombination (if necessary)

2. Relative volume test

3. Differential liberation test

4. Flash separation tests

5. Fluid viscosity

For a gas-condensate system:

1. Fluid compositions and recombination

2. Relative volume test (dewpoint observation)

3. Depletion study

4. Retrograde volume calculations

A.12. Reservoir Fluid Correlations

Reservoir Fluid Correlations

Fluid properties may be determined in several ways:

1. Direct measurement on a sample of the fluid in the laboratory.

2. Calculation from a compositional analysis and knowledge of the properties of the components (additive volumes).

3. Empirical correlations with other physical parameters.

Correlations exist for many properties and new ones have been developed over time. Most correlations are available as graphs, nomographs, tables, or equations. They are useful for making estimates for designing the correct sampling procedure required for a given well and as a check against results obtained in the lab. If sampling is impossible or uneconomic, correlation data may have to suffice for engineering calculations.

Table 1., below, shows the most common reservoir fluid correlations and the researchers responsible for them. We will now briefly review each of those listed.

Desired Property	Correlating Properties	Researcher
*Gas systems*
Pseudo-criticaltemperature and pressureGas viscosity	Molecular weight, gas gravity Molecular weight,temperature, pressure	Mathews, Roland& Katz; KatzCarr, Kobayashi& Burrows
*Gas-condensate systems*
Dewpoint pressure	Oil gravity, temperature,GOR	Sage & Olds
	Oil gravity, temperature,GOR
Well stream gravity	Oil gravity, gasgravity, GOR	Standing
	Oil gravity, gasgravity, GOR
*Oil systems*
Bubble-point pressure	GOR, oil gravity,gas gravity, temperature	Standing; Lasater; Vazquez &Beggs; GlasØ
	GOR, oil gravity,gas gravity, temperature	Standing; Lasater; Vazquez &Beggs; GlasØ
Oil density	Compositional analysis,pressure, temperature	Standing & Katz
	Compositional analysis,pressure, temperature
Solution GOR	Bubble-point pressure, gas gravity, oil gravity; temperature	Standing; Lasater; Vazquez &Beggs; GlasØ
		Standing; Lasater; Vazquez &Beggs; GlasØ

Formation volume factor	GOR, oil gravity,gas gravity, temperature	Standing; Vazquez & Beggs;GlasØ
	GOR, oil gravity,gas gravity, temperature	Standing; Vazquez & Beggs;GlasØ
Isothermal compressibility	Oil gravity, gas gravity, B0, GOR, temperature, pressure	Calhoun; Trube; Vazquez & Beggs
		Calhoun; Trube; Vazquez & Beggs

*Oil viscosity*
Dead oil	Temperature, oil gravity	Beal; Beggs & Robinson;

Gas-saturated oil	Dead oil viscosity, GOR	Chew & Connally; Beggs &

Undersaturated oil	Saturated oil viscosity,	Beal; Vazquez & Beggspressure, bubble-point pressure


*Formation Water*
Formation volume factor	Pressure, temperature	Dodson & Standing

Table 1. Reservoir fluid correlations

Gas Systems

Pseudocritical Temperature and Pressure

When a gas composition is unavailable or incomplete, a determination of the pseudocritical pressure and pseudocritical temperature is possible if the molecular weight or specific gravity of the gas is known. Figure 1 (Pseudocritical temperature and pressure for heptanes-plus fraction) is a correlation for determining the pseudocritical properties of the heptanes-plus fraction usually reported in a gas analysis. This correlation, along with the critical properties of the lighter components, allows the calculation of reduced pressure and temperature and, ultimately, the compressibility factor.

Figure 2 (Pseudocritical constants as functions of gas gravity) is a correlation for pseudocritical properties based upon gas gravity, for use when even a partial analysis is unavailable. Standing (1952) has provided equations for the curves in Figure 2 :

gas:

p_pc = 677 + 15 _g- 37.5_g2

T_pc = 168 + 325_g - 12.5_g2

condensate:

p_pc = 706 - 51.7 _g- 11.1_g2

T_pc = 187 - 330_g - 71.5_g2

Gas Viscosity

Gas viscosity has been correlated using the law of corresponding states. Carr, Kobayashi, and Burrows (1954) have developed a two-step procedure for calculating gas phase viscosity. First, the gas viscosity at atmospheric pressure is computed using Figure 3 (Viscosity of gases at one atmosphere pressure) with the appropriate corrections for nonhydrocarbon constituents. The viscosity at reservoir conditions is then determined from Figure 4 (Viscosity ratio as a function of pseudoreduced temperature and pressure) using reduced pressure and temperature. The pseudocritical properties must be determined from a compositional analysis or from the correlations described earlier.

Gas Condensate Systems

Dewpoint Pressure

Sage and Olds (1947) have correlated dewpoint pressures with stock tank oil gravity, reservoir temperatures, and gas-oil ratio. The results for oil at 100F are given in tabular form in Table 1, below. This correlation is based on data from only five California condensate systems and should be considered a general guideline.

Tank Oil Gravity	Gas-Oil Ratio, Cubic Feet per Barrel @100F
API	15,000	20,000	25,000	30,000	35,000	40,000
52	4,440	4,140	3,880	3,680	3,530	3,420
54	4,190	3,920	3,710	3,540	3,410	3,310
56	3,970	3,730	3,540	3,390	3,280	3,180
58	3,720	3,540	3,380	3,250	3,140	3,060
60	3,460	3,340	3,220	3,100	3,010	2,930
62	3,290	3,190m	3,070	2,970	2,880	2,800
64	3,080	3,010,	2,920	2,840	2,770	2,700

Table 1. Correlation of dewpoint pressure with GOR, temperature and oil gravity for five California condensate systems (Sage and Olds, 1947; courtesy of SPE/AME)

The dewpoint pressure can also be computed using equilibrium ratios, and the selection of a proper K-value for the heavy component fraction is critical. Although there are correlations for estimating the equilibrium ratios of the heavy component fraction based on molecular weight, specific gravity, and boiling point, these methods are usually not accurate enough for dewpoint calculations. Laboratory vapor-liquid data should be obtained.

Well Stream Gravity

Standing (1952) presents a handy relationship for calculating the composite well stream gas gravity from the data available at the surface from a producing gas condensate well. The correlation is given in Figure 1 (Well stream gravity as a function of GOR, gas gravity and oil gravity) and the equation is given in Table 2, below. After determining the well fluid gravity, the earlier correlations for pseudocritical properties can be utilized. If retrograde condensate forms within the reservoir, the surface measurements of GOR and gravity are no longer representative.

where:

T = gravity of total well stream

MW^o = molecular weight of condensate

R_s = gas-oil ratio (SCF/STB)

_g= gas gravity

^o= condensate specific gravity

Table 2. Equation for Standing's correlation given in Figure1 .

Oil System

Bubble-point Pressure

Four possible correlations for bubble-point pressure are presented:

Standing's: based on 105 experimentally determined data points from 22 different California crude oil-gas mixtures.

Lasater's: based on 158 experimentally determined data points from different crude oil-gas systems from Canada, the United States, and South America.

Vazquez and Beggs's: based on 6004 data points divided into two groups according to relative volatility (<30 °API and >30 °API).

GlasØ's: based on data from 45 North Sea oil samples with UOP characterization factors of about 11.9.

The equations for each of these correlations are given in Figure 1 (Charts for Lasater's bubble-point correlation (a) bubble-point factor Pf as a function of the mole fraction of gas in solution Yg and (b) molecular weight related to tank oil gravity) and Table 1, below. Graphical presentations of these correlations are available in their respective references.

Standing

Lasater

where pf and M_oare respectively determined from Figure 1 (a & b).

Vazquez
and Beggs

Coefficient	°API 30	°API > 30
C1	24.64	56.06
C2	1.0937	1.1870
C3	11.172	10.393

_gs= _gp[1.0 + 5.912 10-5 · °API · Tsp · log (psp/114.7)]

GlasØ

p_b = antilog [1.7669 + 1.7447 · log p_b* - 0.30218 · (log p_b*)2]

where:

Table 1.
Bubble-point pressure correlations (Standing, 1947; Lasater, 1958; Vazquez and Beggs, 1980; GlasØ, 1980; courtesy of API, SPE/AIME, JPT. and JPT. respectively)

Oil Density

The additive volume method of calculating liquid density is not applicable when significant amounts of the lighter hydrocarbon components are present. The densities of methane and ethane depend upon the composition of the heavier fraction of the liquid; this is related to the forces of intermolecular attraction. Standing and Katz (1942) have developed correlations of the apparent liquid density of methane and ethane with the density of the system. With these data they developed a modification of the additive volume method that utilizes the chart shown in Figure 2 (Pseudoliquid density of systems containing methane and ethane, at atmospheric pressure and standard temperature) along with a compositional analysis. The density of the pro pane-plus is computed using the additive volume method. The weight percent of the ethane in the ethane-plus system is also computed using the compositional analysis and the molecular weight of each component. Figure 2 is then used to determine the density of the system, including methane and ethane. The density corrections to reservoir conditions are then made using Figure 3 (Density correction for pressure) (pressure correction), followed by Figure 4 (Density correction for temperature) (temperature correction).

If the composition of the liquid is not available, Katz () has developed a correlation that still allows the calculation of the liquid density. The gas gravity and the tank oil gravity are used to obtain an apparent density of dissolved gas from Figure 5 (Apparent density of dissolved gas as a function of gas and oil gravity). The equations given in Table 2, below, are then used to determine the pseudoliquid density at standard conditions, and Figure 3 and Figure 4 are used to correct the value to reservoir conditions.

= (Wg + 5.614 )/[(Wg/) + 5.614]

where:

W_g= weight of gas dissolved in 1.0 STB (lb/STB)

R_s = solution gas-oil ratio (SCF/STB)

_g = gas gravity

_a= pseudo-liquid density (lb/ft3)

_ga = apparent density of dissolved gas (lb/ft3)

_ost= density of stock tank oil (lb/ft3)

Table 2. Calculation of liquid density using Figure 5

Formation Volume Factor

Correlations for the formation volume factor of a saturated crude oil have been developed by Standing (1947), Vazquez and Beggs (1976), and Glas (1980), using the data they accumulated for their bubble-point correlations. The equations are given in Table 3, below.

Standing

Bob = 0.972 + 1.47 10-4

Vazquez and Beggs

Bob = 1.0 + C1·Rs + C2·(T-60)·+C3·Rs (T-60)

Coefficient	°API 30	°API > 30
C1	4.677 10-4	4.670 10-4
C2	1.751 10-5	1.100 10-5
C3	-1.8106 10-8	1.337 10-9

GlasØ

Bob = 1.0+antilog [-6.58511+2.91329 log Bob*-0.27663·(log Bob*)2]

where:

Bob = Rs · + 0.968 · T

Table 3. Correlations for formation volume factor (Standing, 1947; Vazquez and Beggs, 1980; GlasØ, 1980; courtesy of API, JPT, and JPT, respectively)

Solution GOR

Correlations for the solution GOR at the bubble point have also been developed and are presented in Table 4, below.

Standing

Lasater

where yg and are respectively determined from Figure 1 (a &b).

Vazquez and Beggs

antilog (C3 °API/(T + 459.67)]

Coefficient	°API 30	°API > 30
C1	27.64	56.06
C2	1.0937	1.1870
C3	11.172	10.393

GlasØ

Rs =

where:

p_b* = antilog [2.8869 - (14.1811 - 3.3093 · log p_b)0,5]

Table 4. Correlations for solution GOR at the bubble point (Standing, 1947; Lasater, 1958; Vazquez and Beggs, 1980; GlasØ, 1980; courtesy of API, SPE/AIME, JPT, and JPT respectively)

Isothermal Compressibility

Several correlations exist for determining isothermal compressibility of under-saturated oil. Calhoun's (1947), Trube's (1957), and Vazquez and Beggs's (1976) correlations are presented in Table 5, below.

Calhoun

See Figure 10 (Calhoun's correlation of isothermal compressibility with the specific gravity of the oil at bubble point pressure.)

Trube

See Figure 6 (Variation of pseudocritical temperature with specific gravity and bubble point of liquid), Figure 7 (Pseudocritical pressure versus pseudocritical temperature and specific gravity of oil at ) and Figure 8 (Reduced compressibility for undersaturated reservoir oil).

Vazquez and Beggs

co = (-1433.0+5.0·Rs+17.2 T-1180.0 _gs+12.61·°API)/(p 105)

Table 5. Correlations for isothermal compressibility (Calhoun, 1960; Trube, 1957; Vazquez and Beggs, 1980; courtesy of university of Oklahoma Press, SPE/AIME, and JPT, respectively)

Trube's (1957) correlation is based on pseudoreduced critical properties much like the compressibility factor for natural gas ( Figure 8 ). The critical properties are a function of liquid density and the bubble-point pressure of the liquid, both corrected to 60°F. This may be done using the previously mentioned correlations. Figure 9 (Pseudocritical pressure and temperature for undersaturated oil as a function of specific gravity) may be used in place of Figure 6 and Figure 7, if only the liquid density is known.

Calhoun's correlation requires the use of Figure 10 after calculating the specific gravity of the oil at bubble-point pressure.

Beal

where:

a = antilog

Beggs and Robinson

_od
= 10x - 1

where:

X = Y · T-1.163

Y = 10z

Z = 3.0324 - 0.02023 · °API

GlasØ

_od = (3.141 1010) · (log °API)[10.313 · (log T) - 36.447]

Table 6. Correlations for dead oil viscosity (Beal, 1970; Beggs and Robinson, 1975; GlasØ, 1980; courtesy of SPE/AIME, JPT, and JPT respectively)

Chew and Connally

_od
= a · (od)b

where:

a = 0.20 + 0.80 · antilog (-0.00061 · Rs)

b = 0.43 + 0.57 · antilog (-0.00072 · Rs)

Beggs and Robinson

_od
= a (od)b

where:

a = 10.715 · (Rs + 100)-0.515

b = 5.44 · (Rs + 150)-0.338

Table 7.
Correlations for oil viscosity at the bubble-point (Chew and Connally, 1959; Beggs and Robinson, 1975; courtesy of SPE/AIME and JPT respectively)

Beal

_o
= _ob
+ 0.001 · (p - p_b) · (0.024 · ob1.6 + 0.038 · ob0.56)

Vazquez and Beggs

where:

m = 2.6 · p1.187 antilog [(-3.9 10-5) · p - 5.0]

Table 8. Correlations for undersaturated oil viscosity (Beal, 1970; Vazquez and Beggs, 1980; courtesy of SPE/AIME and JPT, respectively)

Oil Viscosity

Oil viscosity is generally determined by first calculating the "dead" oil (no solution gas) viscosity and using that value to obtain the viscosity of the oil at the bubble point. The bubble-point oil viscosity can then be used to determine the undersaturated oil viscosity. Equations for each of these calculations are given inTables 6, 7 and 8.

Formation Water Correlations

Formation Volume Factor

Dodson and Standing (1944) have correlated formation volume factors for pure water and gas-saturated formation brines. Figure 1 (Water formation volume factor for pure water and pure water saturated with natural gas) gives the formation volume factor for pure water with and without solution gas.

Since an increase in water salinity reduces the amount of dissolved gas, and hence the formation volume factor, a correction must be made for dissolved salt content. This is done by multiplying the difference between the pure water and gas saturated volume factors by a correction ratio obtained from Figure 2 (Reduction in gas solubility due to dissolved salts). The resulting value is added to the formation volume factor for pure water to obtain the corrected value for solution gas and salinity.

Choosing Correlations

In most cases several correlations exist for any given parameter. The values calculated are usually not radically different from one another. Each correlation fits the data upon which it was based within a certain percentage error, generally less than 5%. If the oil or gas is representative of the fluids used in the correlation's development, a more accurate estimate of the correlated parameter can be expected. Thus, GlasØ would be the obvious choice for a North Sea oil, while Standing would be better for a California crude. Additionally, some efforts have been made to determine which correlations are most representative of other producing areas. Sutton and Far-shad (1984) have compared each of the major correlations with laboratory data from 31 Gulf of Mexico crude oils. Their findings show that GlasØ's correlation generally yields the best results for bubble-point pressure, solution GOR, and oil formation volume factor, except at GORs greater than 1400 SCF/STB (249 m3/m3) and bubble-point pressures above 7000 psia (48,000 kPa), where Vazquez and Beggs's correlation is more accurate.

Ostermann, Ehlig-Economides, and Owolabi (1983) determined that Alaskan crude oils are best represented by Glas (bubble-point pressure), Standing (oil formation volume factor) and Beggs and Robinson (oil viscosity).

If one or several PVT analyses are available within a producing area, calculations using each correlation can be compared to the actual laboratory data, and the most representative correlation may be chosen.

A.13. References and Additional Information

References

Amyx, J.W., D.M. Bass, and R.L Whiting. 1960. Petroleum reservoir engineering — physical properties. New York: McGraw-Hill.

API 1966. RP44: Recommended Practice for sampling Petroleum reservoir fluids. 1st ed. Dallas: American Petroleum Institute.

API 1981. RP45: Recommended practice for analysis of oil-field waters. 2nd ed. Dallas: American Petroleum Institute.

Arps, J.J. 1962. Estimation of primary oil and gas reserves. In Petroleum Production
Handbook, ed. T.C. Frick. 2:37-1-37-56. SPE-AIME.

Banks, R.F., and P.J. King. 1984. Modern petroleum technology: Part 1. 5th ed. Institute of Petroleum, New York: John Wiley & Sons, 279-327.

Beal, C. 1970. The viscosity of air, water, natural gas, crude oil and its associated gases at oil field temperature and pressure. SPE Reprint Series No. 3 Oil and Gas Property Evaluation and Reserves Estimates. SPE-AIME, 114-127.

Beggs, H.D., and J.R. Robinson. 1975. Estimating the viscosity of crude oil systems. JPT (Sept.) 1140-1141.

Bicher, L.B., and D.E. Katz. 1944. Viscosity of natural gases. Trans. AIME 155.

Brown, G.G., D.L. Katz, G.G. Obertell and R.C. Alien. 1948. Natural gasoline and the volatile hydrocarbons. NGAA (Tulsa).

Burcik, E.J. 1979. Properties of petroleum reservoir fluids. Boston: IHRDC.

Buthod, P. 1962. Properties of crude oil and liquid condensates. Petroleum Production Handbook, ed. TC. Frick. SPE-AIME, 2:18-1-18-23.

Calhoun, J.C., Jr. 1960. Fundamentals of reservoir engineering. rev. ed. Norman, OK: University of Oklahoma Press.

Chew, J. and C.A. Connally, Jr. 1959. A viscosity correlation for gas-saturated crude oils. Trans. AIME, 216:23-25.

Carr N.L., R. Kobayashi, and D.B. Burrows. 1954. Viscosity of hydrocarbon gases under pressure. Trans. AIME 201.

Collins, A.G. 1975. Geochemistry of oilfield waters. New York: Elsevier.

Craft, B.C., and M.F. Hawkins. 1959. Applied petroleum reservoir engineering. Englewood Cliffs, N.J: Prentice-Hall, 149.

Dake, L. P. 1978. Fundamentals of reservoir engineering. New York: Elsevier, 56.

Dodson, C.R., and M.B. Standing. 1944. Pressure-volume-temperature and solubility relations for natural gas-water mixtures. Drilling and Production Practice. API.

Eichelberger, W.C., and A. Michels. 1962. Hydrocarbon water and formation water correlations. In Petroleum production handbook, ed. T.C. Frick. Vol 2: 22-13. SPE-AIME.

Eilerts, C. K., H.A. Carlson, and N.B. Mullens. 1948. Effect of added nitrogen on compressibility of natural gas. World Oil (June and July).

Engineering Data Book. 9th ed. 1972; 5th rev. 1981. Tulsa: Natural Gas Processors Suppliers Association.

Flopetrol. 1979. Guidelines for reservoir fluid sampling. Melun, France: Flopetrol.

GlasØ, O. 1980. Generalized Pressure-volume-temperature correlations. JPT 32 (May):765795.

Katz, D.L., 1942. Prediction of the shrinkage of crude oils. Drilling and Production Practice 137. API.

Katz, D.L., and K.H. Hachmuth. 1937. Vaporization equilibrium constants in a crude oil-natural gas system. Ind. & Eng. Chem. 29:1072.

Lasater, J.A. 1958. Bubble point pressure correlation. Trans. AIME 213:279381.

Levorsen, A.l. 1967. Geology of petroleum. 2nd ed. San Francisco: W.H. Freeman.

Mathews, T.A., C.H. Roland, and D.L. Katz. 1942. High pressure gas measurement. Refiner 21 (June).

McCain, W.D., Jr. 1973. The properties of petroleum fluids. Tulsa: Petroleum Publishers Co.

Noad, D.F. 1962. Water analysis data, interpretation and applications. J Can. Pet Tech. 1:82-89.

Ostermann, R.D., O.O. Owolabi, and C.A. Ehlig-Economides. 1983. Correlations for the reservoir fluid properties of Alaskan crudes. SPE 11703. Paper presented at Ventura, California, Regional Meeting.

Peng Ding-Yu, and D.B. Robinson. 1976. A new two-constant equation of state. I & EC. Fundamentals 15(1):59.

Perry, J.H. 1950. Chemical engineers handbook. 3rd ed. New York: McGraw-Hill.

Redlich, O. and J.N.S. Kwong. 1949. On the thermodynamics of solutions V — an equation of state. Fugacities of gaseous solutions. Chem Reviews 44:233.

Roland, C.H. DE. Smith, and H.H. Kaveler. 1941. Equilibrium constraints for a gas- distillate systems. Oil and Gas J. 39(March 27):128.

Sage B.H., and R.H. Olds. 1947. Volumetric behavior of oil and gas from several San Joaquin Valley fields. Trans. AIME 170.

Schlumberger, Inc. 1982. Reservoir and production fundamentals. Schlumberger, Inc. 3-19.

Standing, M.B. 1952. Volumetric and phase behavior of oil field hydrocarbon systems. New York: Reinhold.

_______ 1947. A pressure-volume-temperature correlation for mixtures of California oils and gases. Drill. and Prod. Prac. (API) 275-287.

Standing, M.B., and D.L. Katz. 1942. Density of crude oils saturated with natural gas. Trans. AIME 146.

Stiff, H.A. 1951. The interpretation of chemical water analysis by means of patterns. JPT 192:376 379.

Sutton, R.P. and F.F. Farshad. 1984. Evaluation of empirically derived PVT properties for Gulf of Mexico crude oils. SPE
13172. Paper presented at the 59th SPE Technical Conference, Houston.

Tickell, F.G. 1921. A method for graphical interpretation of water analysis. 1921 California State Oil and Gas Supervisor Report 6:5-11.

Tissot, B.P., and D.H. Welte. 1978. Petroleum formation and occurrence. New York: SpringerVerlag.

Trube, Albert 5. 1957. Compressibility of under-saturated hydrocarbon reservoir fluids. Trans. AIME. 210:341-344.

Turner, R.G., M.G. Hubbard, and A.E. Dukler. 1969. Analysis and prediction of minimum flow rate for the continuous removal of liquids from gas wells. Trans. AIME. 246:1475-1482.

Van Winger, N. 1950. Viscosity of air, water, natural gas, and crude oil at varying pressures and temperatures. In Secondary Recovery of Oil in the United States. API.

Vazquez, M., and H.D. Beggs. 1980. Correlations for fluid physical property prediction. JPT 32(June):968-1 70.

Recommended Reading

For additional information on the subject of reservoir fluids, fluid sampling, and fluid analysis, the reader is directed to the following excellent texts:

Petroleum Reservoir Engineering: Physical Properties (1960), by J.W. Amyx, D.M. Bass, Jr., and R.E. Whiting, available from McGraw-Hill Inc., New York, NY.

Volumetric and Phase Behavior of Oil Field Hydrocarbon Systems (1981), by M.B. Standing, available from the Society of Petroleum Engineers of AIME, Dallas, TX.

The Properties of Petroleum Fluids (1974), by W.D. McCain, available from PennWell Books, Tulsa, OK.

Several companies specializing in fluid analysis also publish course notes, brochures, and collected papers on their particular procedures and techniques. For example:

A Course in the Phase Behaviour of Hydrocarbon Reservoir Fluids, Reservoir Fluid Analysis Department; Core Laboratories Inc.; Dallas, TX.

Guidelines for Reservoir Fluid Sampling, Flopetrol Johnston; Schlumberger; Melun, France.

Reservoir Fluid Analyses Examples, Weatherly Laboratories; Lafayefte, LA.

The American Petroleum Institute also has published:

Recommended Practice for Sampling Petroleum Reservoir Fluids, API RP44, 1st ed. (January, 1966), available from the API Production Dept., Dallas, TX.

Excellent references on the topic of formation water sampling and analysis are:

Geochemistry of Oilfield Waters, by A.G. Collins (1975), available from Elsevier Scientific Publishing Co., New York.

Recommended Practice for Analysis of Oil-Field Waters, API RP45, 2nd ed. (November, 1968; reissued July, 1981), available from the API Production Dept., Dallas Texas.

Nomenclature

A	=	area
B_g	=	gas formation volume factor, res. bbl/scf (m³/m³)
B_gi	=	B_g at initial conditions
B_ga	=	B_g at abandonment pressure
B₀	=	oil formation volume factor, res. bbl/STB (m³/m³)
B_ob	=	B₀ at the bubble point
B_od	=	differential liberation volume factor
B_odb	=	B_od at the bubble point
B_ofb	=	flash liberation volume factor at the bubble point
B_t	=	total formation volume factor
B_w	=	water formation volume factor, res. bbl/STB (m³/m³)
c	=	compressibility, psia_-1 (kPa_-1)
c₀	=	oil compressibility
c_g	=	gas compressibility
c_w	=	water compressibility
c_pr, c_pR	=	pseudoreduced compressibility
F	=	force
K	=	equilibrium ratio
K_i	=	equilibrium ratio of i_th component
L	=	moles of material in liquid phase
M	=	molecular weight
M₀, MW₀	=	molecular weight of oil
m	=	weight of a quantity of gas
n	=	moles of gas
p	=	pressure, Psia (kPa)
p_i	=	partial pressure of i^th component in a mixture
p_vi	=	vapor pressure of i^th component at temperature T
p_b	=	bubble-point pressure
p_c	=	critical pressure
p_pc	=	pseudocritical pressure
p_R, p_pR, p_pr	=	pseudoreduced pressure
p_sc	=	standard pressure, 14.7 psia (101 kPa)
p_st	=	stock tank pressure
p_sep	=	separator pressure
p_wf	=	flowing bottomhole pressure
	=	average reservoir pressure
p_f	=	bubble-point factor (Lasater)
p_abn	=	abandonment pressure
q_g	=	gas flow rate
R	=	gas constant 10.732 psi ft³ (lb mole)^-1 (°R)^-1 8.3143 J (mol)^-1 (K)^-1
Rs	=	solution gas-oil ratio, scf/STB (sm³/m³)
R_si	=	initial gas-oil ratio
R_sfb	=	flash liberation gas-oil ratio at the bubble point
R_adb	=	R_sd at the bubble point
R_sd	=	differential liberation gas-oil ratio
T	=	temperature, °F, °R, (K)
T_ac	=	standard temperature, 60 °F, 520 °R (289K)
T_sep	=	separator temperature
T_st	=	stock tank temperature
T_R	=	reservoir temperature
T_pr, Ts	=	pseudoreduced temperature
T_c	=	critical temperature, °R
T_pc	=	pseudocritical temperature
V	=	moles of material in gas phase
V	=	volume
Vt	=	total volume
V_b	=	volume at the bubble point
V_R	=	volume at reservoir conditions
V_sc	=	volume at standard conditions
	=	velocity
_g	=	gas velocity
v	=	kinematic viscosity
W_g	=	weight of gas
x_i	=	mole fraction of i^th component in liquid phase
y	=	distance
y_i	=	mole fraction of i^th component in gas phase
y_g	=	mole fraction of gas in solution (Lasater)
z	=	compressibility factor
z_i	=	mole fraction of i^th component in mixture
z_sc	=	compressibility factor at standard conditions (1.0)
_g	=	gas gravity (air = 1.0)
_o	=	oil gravity (water = 1.0)
T	=	gravity of total well stream
_gp	=	gas gravity of separator conditions (Vazquez and Beggs)
_ga	=	separator gas gravity at psep = 100 psig (Vazquez and Beggs)
_ob	=	specific gravity of oil at the bubble point
	=	density, lb/ft3 (kg/m3)
_	=	oil density
_g	=	gas density
_a	=	pseudoliquid density
_ga	=	apparent density of dissolved gas
_ost	=	density of stock tank oil
µ	=	viscosity, cp
µ_od	=	viscosity of "dead" oil
µ_ob	=	viscosity of oil at the bubble point
µ_o	=	oil viscosity
µs	=	gas viscosity

petroleum, crude oil