January 22-23, 2014New Orleans, Louisiana

2

Section | A

Founding Contributors

Exhibitors

Board Members

Section | B

Page 1 of 15

THE PHYSICS AND CHEMISTRY OF NATURAL GAS SAMPLING AND CONDITIONING

Darin L. George, Ph.D., Southwest Research Institute®

James N. Witte, Southwest Research Institute®

Abstract

Industry research has led to changes in industry standards for natural gas sampling, such as the American Petroleum Institute (API) Manual of Petroleum Measurement Standards (MPMS), Chapter 14.1 and Gas Processors Association (GPA) Standard 2166. To best apply these standards, users should understand the physical phenomena that can lead to inaccurate samples. This paper will review the physics of natural gas sampling, including phenomena such as adsorption and desorption, vapor-liquid equilibrium, Joule-Thomson cooling, the use of the phase diagram as a sampling tool, and results of research on the physics of natural gas sampling that have led to developments in industry standards.

Introduction

Accurate compositional analysis and gas quality determination is of high importance to the natural gas industry. For instance, accurate data on the water vapor content of natural gas streams is needed to identify potentially corrosive operating environments before significant damage to natural gas pipelines can occur. Hydrocarbon dew points (HCDPs) determined from gas analyses are used as measures of the quality of a natural gas stream and as one criterion for assessing compliance with transportation tariffs. Accurate gas quality data is also crucial to the effective introduction of new gas supplies, such as shale gas, biogas, and land fill gas, into the natural gas transmission grid and is crucial to its efficient use by consumers.

Natural gas samples must be representative of the true composition of a flowing gas stream to avoid unnecessary interruptions of gas supplies that may, in reality, meet safety and tariff requirements, or to avoid allowing poor quality gas entry into a transmission or distribution system. Research funded by several industry organizations has identified several physical and chemical causes of distorted sample compositions and has led to improved techniques for sampling natural gas streams for hydrocarbon and moisture content. Many of these improvements have been documented in industry standards for natural gas sampling, such as API MPMS, Chapter 14.1 (2006) and GPA Standard 2166 (2005).

It should be remembered that these documented procedures are not “cookbook approaches,” but guidelines to avoid sample distortion. To best apply the methods described in these standards, those who design and use sampling equipment must understand the physical phenomena that can lead to inaccurate samples. To this end, this paper reviews the physics of natural gas sampling, including phenomena such as adsorption and desorption, vapor-liquid equilibrium, and Joule-Thomson cooling. The paper will also explain the use of the phase diagram as a sample system design tool, and briefly discuss recent research on the physics of natural gas sampling that have led to these new industry standards.

Representative Samples

The objective of natural gas sampling is to obtain a sample that is representative of the bulk characteristics of the flowing gas stream. Spot sampling will only yield a sample that is representative of the composition at that particular instant.

Long-term proportional-to-flow sampling methods, such as automatic composite samplers and real time sampling on-line gas chromatographs, will give analytical results that are significantly more representative over time than spot samples. However, the equipment for both types of sampling systems can be fouled and made inoperable by liquid contamination.

It is difficult to accurately apply the term “representative” when one considers the occurrence of gas condensation within a pipeline. Given that the objective is to sample the gas and perform a compositional analysis of that sample, when components of the gas begin to condense they have a tendency to drop out of the vapor phase and adhere or adsorb to the pipe wall (Nored and George, 2003 and George and Kelner, 2010). In this condition, a fraction of the gas has now been removed from the flowing stream and may not be available for delivery to the sample probe. This can lead to a gas sample that has a reduced fraction of “heavy” hydrocarbon components, such as pentanes, hexanes, etc.

Page | 1

Page 2 of 15

As the condensation process continues the droplets begin to stream along the walls of the pipe. This can lead to the pooling of liquids at low velocity points within the pipe, or possibly the delivery of droplets to the tip of the sample probe. These droplets are referred to as “free” liquids. The inclusion of free liquids in the sample may bias the sample, causing it to be “richer” than the original gas composition because of the inclusion of a disproportionate fraction of heavy hydrocarbon components. This biasing may affect the gas heating value, as well as the calculated hydrocarbon dew point and gas density.

Gas Components Dissolved into Liquids

Certain “free” liquids can dissolve significant amounts of natural gas sample components out of the gas phase. These liquids may find their way into the sampling system as residue from previous sample streams or as solvents left behind from cleaning procedures. As a rule, liquid residue with the same or similar chemical composition as components in gas samples will dissolve those components from the samples and distort the sample compositions (Behring and Kelner, 1999).

An example of this for hydrocarbons can be seen in the test results shown in Figure 1. Samples of a rich natural gas were placed in constant-volume (CV) and constant-pressure (CP) sample cylinders containing several liquid residues. The liquids included water and glycol left over from displacement sampling procedures, a liquid hydrocarbon (HC) mixture of n-paraffin hydrocarbons and SAE-30 compressor oil, and DuPont Krytox® lubricant. After storage in the cylinders for two to three days, the samples were then analyzed to determine if components had dissolved into the contaminants and affected the sample compositions. The water and glycol liquids did not absorb heavy hydrocarbon components from the samples, and caused no notable distortion of the sample heating value or density. In the tests with liquid hydrocarbon residues, the liquids did absorb hydrocarbon components from the gas, lowering the sample density and heating value by as much as 8%. In the test with Krytox lubricant, it was found that the cylinder seals had previously been exposed to liquid hydrocarbons. This residue was responsible for the sample distortion, rather than the Krytox.

As with adsorption/desorption, the amount of a particular component dissolved in the liquid residue will depend on pressure and temperature conditions of the sample in the container and the amount of the component present in the sample. Sample distortion also is possible where the liquids are in a sample line carrying flowing gas. In this case, when the stream pressure and temperature change, the flowing sample will undergo a similar change in composition to that for adsorption/desorption, as the equilibrium between component levels in the stream and in the liquid changes.

Figure 1. Effect of Residual Liquids in Sample Cylinders on the Density and Heating Value of 1,500 Btu/scf Natural Gas Samples (Behring and Kelner, 1999). (In the test with Krytox lubricant, the cylinder seals had previously been exposed to liquid hydrocarbons, which were responsible for the sample distortion.)

Page | 2

Page 3 of 15

Vapor-Liquid Equilibrium

The causes of sample distortion discussed to this point have all involved components of the gas sample coming into contact with another substance, such as equipment surfaces or a liquid contaminant. When causes of adsorption/desorption and dissolution/elution have been minimized, sample distortion can still occur simply due to changes in the physical state of the gas sample itself. This source of distortion is related to the properties of the components in the sample, and under the right conditions, the tendency of certain components (such as water vapor or heavy hydrocarbons) to undergo phase change and condense out of the gas sample. To explain this mechanism, we must first define some key concepts.

Consider a closed vessel that contains a pure substance, with some number of molecules of the substance in the liquid phase and the rest in the gas phase (Figure 2). The amount in each phase is not critical to the discussion. Even if the substance is in a steady state, some small number of molecules of the substance will move back and forth between the liquid and gaseous phases. At a constant temperature and pressure inside the vessel, the number of vapor molecules condensing to liquid equals the number of liquid molecules evaporating to the gas phase. This system is said to be in vapor-liquid equilibrium, since the amounts in each phase remain constant over time. The pressure that the vapor phase of the pure substance exerts on the walls of the vessel is known as its vapor pressure.

The relative amount of molecules in each phase depends on the pressure and temperature of the system. Suppose we increase the temperature inside the vessel (Figure 3). This will add energy to the molecules in the vessel, and more molecules will tend to escape from the liquid phase into the vapor state. A new vapor-liquid equilibrium will be reached with a larger number of molecules in the vapor phase, and a larger number of molecules moving back and forth between phases. Once this dynamic equilibrium has been reached, the vapor phase will reach a new, higher vapor pressure. This is because the vapor pressure of a substance is proportional to the relative amount of molecules of that substance in the vapor phase, a principle known as Raoult’s law (Bailar, et al., 1978).

Figure 2. Example of a Substance in a Closed Vessel in Vapor-liquid Equilibrium (Mayeaux, 2006)

Figure 3. Example of a Change in Vapor-liquid Equilibrium for a Pure Substance with Increasing Temperature (Mayeaux, 2006)

Page | 3

Page 4 of 15

Now consider a mixture containing multiple components, such as natural gas. At a constant temperature and pressure, we will again see vapor-liquid equilibrium, as the number of molecules evaporating from the liquid equals the number of molecules condensing from the gas. Raoult’s law also extends to mixtures containing more than one component, so that the pressure of the vapor phase is related to the number of molecules in the vapor phase. However, the proportionality between the number of molecules in the liquid phase and the vapor phase will not be the same for each component. Even with the mixture at a uniform temperature, one component may have 50% of its molecules in the vapor phase, while another component may have only 10% of its molecules in the vapor phase. If there are equal amounts of each component in the liquid phase, there will almost never be equal amounts of each component in the gas phase over the liquid. Since the gas phase has different amounts of each component contributing to the overall vapor pressure, we say that each component has a different partial pressure. In natural gas mixtures, the “lighter” components (e.g., methane, ethane, etc.) will have higher partial pressures than the “heavier” components (e.g., nonane, decane, etc.) (see Figure 4).

Figure 4. Example of Different Partial Pressures for Different Components in a Multi-component Mixture at Vapor-liquid Equilibrium (Mayeaux, 2006)

Now we extend the discussion of pressure and temperature changes to the case of this multi-component mixture. Because each component has its own equilibrium relationship between amounts in the liquid and vapor phase, pressure and temperature changes will affect each component differently. Lowering the temperature of the mixture, for example, will drive more of the heavier, less volatile molecules into the liquid phase (Figure 5). Since the various components do not condense equally, the temperature drop will change the composition of the gas phase. Similar changes in the gas composition will occur with increases in temperature, or with changes in the total pressure of the gas phase.

Figure 5. Example of a Change in the Gas Phase Composition of a Multi-component Mixture with a Drop in Temperature (Mayeaux, 2006)

Page | 4

Page 5 of 15

This key concept – that the composition of a gas mixture changes when changes in pressure or temperature affect its equilibrium with a liquid phase – must be kept in mind when avoiding sample distortion.

Phase Changes and the Phase Diagram

While the previous section dealt with equilibrium conditions between gas and liquid, a natural gas pipeline will ideally carry only gas. If the stream to be sampled is completely in the gas phase, then the pressure and temperature of each component in the gas phase is such that it would be completely gaseous if it were a pure substance, not part of a mixture, at the same temperature and pressure. A sufficient change in pressure and temperature during sampling, however, can cause specific components to condense into liquid, changing the composition of the gas phase and potentially distorting the sample composition.

As explained by George and Kelner (2006), the dew point is defined as the pressure and temperature at which specific constituents in a natural gas mixture begin to change phase. For instance, if the temperature of a natural gas mixture is reduced while the pressure remains constant, the temperature at which hydrocarbons begin to condense from the gas phase to the liquid phase is the hydrocarbon dew point temperature. If the pressure of a natural gas is increased while the temperature remains constant, the pressure at which hydrocarbon condensation begins is the hydrocarbon dew point pressure. Similarly, a temperature and pressure condition at which water vapor begins to condense from the mixture is called the water vapor dew point. Note that very rarely is the hydrocarbon dew point the same as the water vapor dew point for any given natural gas mixture.

These two types of dew points follow different trends with temperature and pressure. When plotted on a graph of pressure versus temperature, the water vapor dew point curve follows a simple curve, with the dew point pressure increasing with increasing temperature. The hydrocarbon dew point curve is much more complex, but is often considered more crucial to natural gas sample integrity during sampling processes. Indeed, the hydrocarbon dew point is perhaps the single most important property to consider in natural gas sampling. If the sample temperature drops below the hydrocarbon dew point temperature, a significant loss in hydrocarbon content can occur, resulting in errors in volumetric flow rate, heating value, and other calculated gas properties.

A phase diagram, such as the hydrocarbon phase diagram shown in Figure 6, describes the phase change behavior of a given natural gas mixture. It will be used here to illustrate the effect of natural gas sampling processes on natural gas. Line A-B in Figure 6 is the bubble point curve. The bubble point is reached when an infinitesimal amount of gas appears during a decrease in pressure of a liquid hydrocarbon mixture at constant temperature. Line B-C-D-E is the dew point curve. It represents the range of pressures and temperatures at which gas/liquid phase changes occur with a natural gas mixture. Line D-E is the lower, or normal, dew point curve. Condensation associated with the conditions defined by this curve may occur during a pressure increase, such as when compressing a gas sample from a vacuum gathering system into a sample cylinder, or during a temperature decrease, such as occurs when the contents of a sample cylinder are exposed to cold ambient temperatures.

Figure 6. Typical Natural Gas Phase Diagram (George and Kelner, 2006)

Temperature (F)

-200-180-160-140-120-100-80 -60 -40 -20 0 20 40 60 80 100 120 140

Pre

ssure

(psia

)

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

Gas

2 - PhaseRegion

(Vapor/Liquid)

CompositionN

2 = 2.05533

CO2

= 0.5132

C1 = 82.6882C2 = 6.9665C3 = 4.5441i-C4 = 1.1559n-C4 = 1.2856i-C5 = 0.3983 nC5 = 0.2624C6 = 0.0836C7 = 0.0273C8 = 0.0167C9 = 0.0009

MaximumTemp.

MaximumPressureLiquid

CriticalRegion

Bubble Point Curve(Line A-B)

Retrograde Dew Point Curve(Line C-D)

Normal Dew Point Curve(Line D-E)

A

E

D

C

B

Dew Point Curve(Line B-E)

Page | 5

Page 6 of 15

Retrograde condensation is a phenomenon that occurs in many common natural gas mixtures. It is characterized by the presence of two hydrocarbon dew points at a given pressure or temperature. Retrograde condensation can occur during temperature increases at constant pressure or during pressure reductions at constant temperature. The points along line B-C-D represent the pressures and temperatures at which retrograde condensation occurs. For example, in Figure 6, a temperature increase at a constant pressure of 1,400 psi from -60°F to -30°F will cross the dew point curve and cause gaseous components to condense. Similarly, a pressure reduction from 1,500 psia to 1,100 psia at a constant temperature of 40°F will cross the retrograde dew point curve and lead to condensation. Retrograde condensation is a characteristic of natural gases that should be considered both when sampling a natural gas stream and when designing gas sampling systems, since the range of pressures and temperatures of the retrograde dew point curve can be encountered during common sampling processes.

Tests conducted with natural gas sampling methods have shown that allowing the gas sample to drop below the hydrocarbon dew point temperature will distort the sample composition and lead to errors in properties calculated from sample analyses, such as heating value and density. The phase diagram shown in Figure 7 illustrates how different processes common to natural gas sampling can cause the temperature of the sampled gas to fall below the hydrocarbon dew point. Path 1-2 represents the Joule-Thomson cooling process that occurs when natural gas flows through a regulator or partially closed valve and undergoes a drop in pressure. Condensation and sample distortion can occur during this “throttling” process, which will be described later in this paper. The cooling can be offset through the application of sufficient heat to the sampling system, as shown by path 1-3. Path 4-5 shows how condensation of a sample can occur if the sample container and its contents are exposed to an ambient temperature below the hydrocarbon dew point temperature.

Figure 7. Natural Gas Phase Diagram Showing Several Common Processes in Natural Gas Sampling that Can Cause Condensation and Gas Sample Distortion (George and Kelner, 2006)

At this point, a simple example will serve to illustrate the potential for errors due to phase change of a sample. Assume that a representative sample of a natural gas stream with the compositional makeup listed in Table 1 is captured in a standard 300 cc constant-volume sample cylinder. Suppose that the sample is collected at a pressure of 75 psia and at a temperature above its hydrocarbon dew point of 91°F. Then suppose that the cylinder and its contents are exposed to an ambient temperature of 41°F, well below the hydrocarbon dew point (path 4-5 in Figure 7). As the temperature is reduced below the hydrocarbon dew point temperature, hydrocarbon constituents condense, with heaviest components preferentially condensing first. This condensation causes a decrease in the vapor fraction of the mixture and a corresponding decrease in the heating value of the vapor phase.

Temperature (F)

-200-180-160-140-120-100 -80 -60 -40 -20 0 20 40 60 80 100 120 140

Pre

ssure

(psia

)

0

100

200

300

400

500

600

700

800

900

1000

1100

1200

1300

1400

1500

1600

1700

1800

1900

2000

Gas

2 - PhaseRegion

CompositionN

2 = 2.05533

CO2

= 0.5132

C1 = 82.6882C2 = 6.9665C3 = 4.5441i-C4 = 1.1559n-C4 = 1.2856i-C5 = 0.3983 nC5 = 0.2624C6 = 0.0836C7 = 0.0273C8 = 0.0167C9 = 0.0009

Liquid

CriticalRegion

1

2

Path 1-2: Retrograde condensation during throttling to a lower pressure.

No, or not enough heat tracing on sample line.Path 1-3: Heat tracing can offset the coolingwhich occurs during the throttling process,

thereby avoiding condensation.

4

5 Condensationthat occurs when a sampleor calibration standard is

exposed to ambient temperaturesbelow the hydrocarbon dew point

temperature.

3

Page | 6

Page 7 of 15

Figure 8 shows the potential effect of 41°F gas sampling equipment on the 1,500 Btu/scf natural gas composition listed in Table 1. The horizontal axis shows the temperature of the gas-liquid mixture. The vertical axis on the left shows the vapor (gas) fraction on a molar basis. The liquid fraction is simply one minus the vapor fraction. The vertical axis on the right shows the change in vapor fraction heating value, in Btu/scf, as liquid condenses from the gas sample. Because the heavier components condense first, Figure 8 shows that a small amount of liquid condensation is associated with a large decrease in heating value. At a temperature of 41°F, the loss in heating value amounts to over 70 Btu/scf.

Figure 8. Change in Vapor Fraction and Gas Phase Heating Value Associated with Condensation of a 1,500 Btu/scf Natural Gas Mixture (George and Kelner, 2006) (0.834% C6+ at 75 psia)

Table 1. 1,500 Btu/scf Natural Gas Mixture Used in the Example of Figure 8

Component Mole percent

Methane 64.107

Ethane 10.33

Propane 7.128

Isobutane 2.174

n-butane 6.386

Isopentane 1.874

n-pentane 2.307

Hexane 0.538

Heptane 0.187

Octane 0.086

Nonane 0.023

Decane 0.016

Nitrogen 3.939

CO2 0.906

Total 100.001

Page | 7

Page 8 of 15

It should be noted that this is a simple example of the potential problems caused by distortion of a natural gas sample. In practice, the effect of a distorted gas sample on calculated gas properties is very difficult to predict. The effects of poor sampling technique on gas samples taken under actual laboratory and field conditions are far more complicated and cannot be accurately predicted using current technology.

Physical and Chemical Phenomena Affecting Sample Accuracy

Natural gas streams are routinely analyzed for heavy hydrocarbons, water vapor, hydrocarbon dew points, diluents, sulfur-containing compounds, and other components. The equipment used to analyze the stream can range from on-site manual devices, such as chilled mirror dew point analyzers, to automated devices, such as moisture analyzers and gas chromatographs, to sample cylinders used to transport a sample to an offsite laboratory. All of these situations have something in common: a representative sample must be extracted from the pipeline and sent through a sample line to a sample container or analyzer. Accurate natural gas sampling requires that the sample removed from the pipeline and analyzed represent the true composition of the flowing gas stream.

Research by Behring and Kelner (1999) listed several fundamental physical causes of gas sample distortion. They note that natural gas is not a pure substance, but a composite mixture of organic and inorganic pure gases. When certain components are preferentially depleted from the sample gas, the integrity of the mixture is compromised, and the sample becomes distorted. Depending on the physical mechanism, it may be the heavier hydrocarbon components, water vapor, or sulfur compounds that are removed from or added to the sample. This section describes several physical mechanisms that can distort gas samples. The next section will describe sampling techniques that can avoid distortion by these mechanisms.

Adsorption and Desorption of Gas Components at Surfaces

The first mechanism discussed here is adsorption. This is defined as the attraction and “sticking” of gases or liquids to solid surfaces, either through chemical or physical processes (Weast, et al., 1985). The reverse process, the release of gas or liquid molecules from a solid surface, is desorption. Note that adsorption is not to be confused with absorption, which is the penetration of a gas or liquid into another body, such as water into the pores of a sponge. Figure 9 shows a simple example of adsorption and desorption of molecules at a solid wall. Often, this adsorbed layer of gas phase molecules is no more than one molecule thick (Gregg, 1961). Adsorption and desorption rates will depend on the type of molecule(s) in the gas stream and on other variables that will be discussed shortly.

Figure 9. Adsorption and Desorption of Different Molecules at the Wall of a Sample Tube (Mayeaux, 2006)

Two very different kinds of adsorption are of concern in natural gas sample collection – chemical adsorption and physical adsorption. Chemical adsorption, a chemical reaction between solid surface molecules and gas phase molecules, may not be easy to reverse and can lead to permanent sample distortion. Fortunately, chemical adsorption can be avoided by selecting the proper materials for solid surfaces (e.g., tubing, fittings, containers, coatings, etc.) that will not chemically react with gas phase sample molecules. Stainless steel is a common choice for sampling equipment for this reason.

Physical adsorption, however, is a much more persistent problem. Figure 10 shows the response of a real-time moisture analyzer to a sudden increase in the moisture content of the sample stream. In this test (Barajas and George, 2006), an automated moisture analyzer drew gas samples from a pipeline through a stainless steel sample line containing a heated regulator and heat tracing. All equipment in the sample line was heated to 40°F above ambient temperature, and the sample flow rate to the analyzer was at its maximum.

After a stream of moisture-saturated gas was introduced at the sample point, the analyzer required 45 minutes to register the true moisture content of the saturated gas stream. The delay was caused by water vapor molecules

Page | 8

Page 9 of 15

leaving the stream and adsorbing to the inside walls of the sampling equipment. Because of this, the first sample volume of gas from the once-saturated stream lost some of its moisture content, and the analyzer registered a moisture level less than the true saturated value of the stream. As the walls of the sampling equipment gather more and more water vapor molecules, less and less moisture was drawn from the stream. The moisture adsorbed on the sampling hardware eventually reaches equilibrium with the moisture in the saturated stream, and the analyzer eventually measured the true moisture content of the saturated supply stream.

The process of desorption can also cause errors in sample analysis. In tests where the moisture content of the sample stream changed from saturation to typical custody transfer levels (i.e., less than 7 lbm/MMscf), equilibrium by desorption took significantly longer than in cases where the moisture level increased. As before, water vapor molecules clinging to the sample equipment must reach a new equilibrium with the moisture content of the sample stream. Adsorption and desorption will work to bring about this equilibrium, but stable, accurate measurements will only occur after equilibrium is reached.

The example above discusses adsorption and desorption of water vapor molecules, but heavy hydrocarbons and other components of a gas stream are subject to this process also. Thermodynamic conditions influence this equilibrium for any component, since gas molecules have a higher tendency to physically adsorb to solid surfaces at low temperatures and high pressures. Changing the sample stream temperature and/or pressure will tend to shift the equilibrium and adsorb/desorb gas phase molecules until a new equilibrium condition is reached. Similarly, changes in concentration of a component in the gas stream will cause molecules of the same component on surfaces to adsorb/desorb until a new equilibrium is reached between molecules on solid surfaces and the same molecules in the gas stream.

Figure 10. Response of a Real-time Moisture Analyzer to the Introduction of a Moisture-saturated Sample Stream (Barajas and George, 2006) (Regulator heated to 110°F. A flow rate of 2.1 scfh was maintained.)

Flowing Gas Dynamics and Joule-Thomson Cooling

As the previous section shows, natural gas samples obtained through sample lines or captured in sample cylinders can undergo state changes (changes in temperature and pressure) that can lead to distortion of the gas sample composition. In the case of a sample cylinder exposed to temperatures below the hydrocarbon dew point, the condensed components will remain in the sample cylinder with the remaining gas phase, and the overall contents of the cylinder itself will remain unchanged as long as the cylinder valves are not opened. Research (Behring and Kelner, 1999) has shown that reheating the cylinder above its dew point for an adequate length of time will restore the gas sample to its original composition.

0

5

10

15

20

25

30

35

40

45

50

8:00 8:10 8:20 8:30 8:40 8:50 9:00

Time (Hr:Min)

Mo

istu

re L

ev

el

(lb

m/m

ms

cf)

Test Stream

Reference Stream

8:05 - Switch to Saturated Stream

Gas Saver Off

December 15, 2005

Page | 9

Page 10 of 15

When the temperature of a sample stream drops below its dew point for whatever reason during the collection process – at some location within the sampling system, before it reaches the analyzer or sample container – the integrity of the sample is much more likely to be distorted beyond recovery. This is due to the fact that condensed components (water vapor or hydrocarbons) may collect at points within the sample line, regulator, or other system components and not make it to the analyzer or sample container. The remaining gaseous components will not be representative of the full composition of the original stream being sampled. In this section, two causes of temperature drops within the sample flow will be described.

As discussed by Behring and Kelner (1999), when natural gas flows through a restriction, such as a valve restriction, in the sample line, changes in the state of the gas can occur. If no heat is transferred to or from the gas, if the gas does no work (that is, the gas does not turn a compressor or turbine), and if the gas does not undergo a significant change in elevation, then the first law of thermodynamics states that the total energy of the gas will remain the same (Van Wylen, et al., 1994). However, the flow obstruction may cause the energy within the gas to change from one form to another. In particular, the energy of the gas per unit mass will be redistributed between the static enthalpy per unit mass, h, and kinetic energy per unit mass, V2/2, as shown by the following simplified form of the first law:

(1)

The sum of static enthalpy and kinetic energy is known as the stagnation enthalpy, ho, so called because if the gas were to become stagnant and reach zero velocity, this equation would reduce to h = ho. Note that the enthalpy, h, is a thermodynamic property of the gas, and, therefore, is related to its temperature and pressure. As the gas stream accelerates through the flow restriction, its velocity will increase, and because total energy must remain constant, its static enthalpy, h, will decrease. The decrease in static enthalpy will, in general, cause the gas temperature and pressure to decrease.

Consider an example of gas flowing through a valve restriction, such as a short section of small-diameter sample tubing or a reduced-port valve (see Figure 11). Under steady flow conditions, the gas must accelerate from state 1 into the restriction. At or near the throat of the restriction, i.e., state 2, the gas will attain its greatest kinetic energy, and by Equation (1), its static enthalpy will simultaneously drop. The gas will then decelerate back into the full-diameter tubing, and the static enthalpy will increase again.

Figure 11. Relative Changes in Static Enthalpy and Kinetic Energy of Gas Flowing through a Restriction with No Heat Transfer or Work to or from the Gas (adapted from Behring and Kelner, 1999). (The red line shows how the potential for condensation of heavy components changes as static enthalpy changes.)

If the kinetic energy at state 3 is the same as the kinetic energy of the gas at state 1, then the static enthalpy of the gas will return to its original value. The gas pressure, however, will be lower at state 3 than at state 1 due to frictional losses incurred in passing through the restriction, and a corresponding drop in temperature will occur between states 1 and 3. This process, in which gas flowing through a throttle undergoes a temperature and

o

out

out

in

inh

Vh

Vh ==+=+ enthalpy stagnation

22

22

Page | 10

Page 11 of 15

pressure reduction without performing heat transfer or work, is called the Joule-Thomson process (Van Wylen, et al., 1994). The Joule-Thomson coefficient describes the amount of cooling of the gas per unit pressure drop; a rule-of-thumb value used for natural gas flows is approximately 7°F per 100 psi of pressure drop.

Figure 12 shows a pressure-temperature path (1-2-3) that the gas might follow through the phase diagram in the case where little or no pressure recovery is attained downstream of the flow restriction. Note that the gas may cross the hydrocarbon dew point boundary as it travels in and out of the restriction, depending on the initial thermodynamic state, the amount of acceleration, and so on. This is the region in which condensation of heavy hydrocarbons can occur. If no heat is lost or gained by the gas inside the tubing, and it decelerates back to its original kinetic energy level, then the process will return to a line of constant enthalpy (between points 1 and 3) parallel to the lines shown in the figure. However, the potential for sample distortion still exists due to the acceleration of the gas and the temperature drop to state 2 at the restriction. The red line in Figure 11 shows how the potential for condensation and sample distortion increases within the valve and downstream.

Use of the Phase Diagram as a Sampling Tool

Clearly, it is crucial that a natural gas sample stream be kept above its dew point during sampling to avoid condensation of components from the sample. Dew points can be measured directly in the field using a Bureau of Mines chilled mirror dew point tester (American Society of Testing and Materials (ASTM), 2000), but if the location is remote or measurements at many different sites are needed, use of the tester may not be practical. On the other hand, if reliable data on the gas composition – such as an analysis from an on-site gas chromatograph (GC) – is available before sampling equipment is installed, the dew point curve can be estimated using any of a number of commercial equation-of-state software packages, and the sampling equipment can be designed to maintain representative samples.

Figure 12 gives an example of how a phase diagram can serve as a tool to help avoid sample distortion. Information on sample flow rates, temperatures, and pressures at the entrance to a sampling system would be determined as part of the design process. Thermodynamic software would be used to determine changes in temperature and pressure of the gas as it passes through regulators, reduced-port valves, or sample lines exposed to atmospheric conditions. At critical locations within the sampling system, the temperature and pressure state of the sample would be plotted on the phase diagram to identify any locations where condensation and phase change might occur. The sampling equipment could then be designed to avoid these problems, for example, by heating the sampling equipment to compensate for temperature drops or by reducing pressure drops through regulators.

Figure 12. Example of the Changes in State of a Gas Passing through the Flow Restriction in Figure 11 (Behring and Kelner, 1999)

Any of a number of commercial equation-of-state software packages can be used to predict the hydrocarbon dew point curve of an expected natural gas mixture. However, the prediction is only as good as the accuracy of the composition used as input. Small amounts of heavy hydrocarbons, i.e., n-hexane and heavier, can strongly affect

Page | 11

Page 12 of 15

the hydrocarbon dew point of a gas. Unfortunately, many field GCs cannot identify these heavier components separately, and only report a “lumped C6+ fraction” for the gas being analyzed. Research is ongoing to determine the best methods of “characterizing” the heavier components in this fraction when only the C6+ total is known.

Figure 13 shows several phase diagrams computed for a single gas composition assuming different characterizations for the hexanes and heavier hydrocarbons (George, et al., 2005). The results were compared to experimental hydrocarbon dew point data for the same gas composition to determine the potential errors due to poor characterizations. The worst case was obtained by treating the lumped C6+ fraction as 100% normal hexane. Using this characterization gave computed dew points as much as 35°F below the experimental data. If this characterization had been used to design a sampling system, the samples could cool well below the hydrocarbon dew point, condensing out heavy components and distorting the sample. In general, research has indicated that a natural gas composition must be known through nonane for its dew point to be computed reasonably accurately. This information can be obtained by analyzing the stream using a GC capable of detecting hexanes and heavier hydrocarbons separately. This requires a sample to be carefully obtained and sent offsite for analysis.

Figure 13. Effect of Using Various C6+ Characterizations on Predicted Hydrocarbon Dew Point Curves (George and Kelner, 2006)

Recent Research on Natural Gas Sampling Methods and Methods of Avoiding Sample Distortion

During preparation of the 6th edition of API MPMS, Chapter 14.1, updates were made in several specific areas related to accurate natural gas sampling. These updates were based on results of research performed in support of the standard (Behring and Kelner, 1999; API, 2006), and on recent revisions to other relevant standards. Other research in the public domain has identified other techniques for minimizing sample distortion (Barajas and George, 2006; George and Kelner, 2006). This section will review sampling equipment designs and sampling methods that can reduce the potential for sample distortion through the mechanisms listed above.

Chemical adsorption can be avoided by using materials in sampling systems, such as stainless steel, that are chemically inert. Adsorption of hydrocarbon components onto seal materials can cause leakage and/or seal failure. Viton is a common material for O-rings and other seals, as it minimizes heavy hydrocarbon adsorption. API MPMS, Chapter 14.1 recommends that sample cylinders used in sour and/or corrosive gas service should be specially lined or coated with epoxy, or otherwise be passivated. Very reactive components, such as hydrogen sulfide (H2S), should be analyzed on-site when practical, rather than taken by sample cylinder for offsite analysis.

Page | 12

Page 13 of 15

Physical adsorption cannot be entirely eliminated from sampling systems, but as documented by Behring and Kelner (1999) and Barajas and George (2006), several steps can be taken to minimize sample distortion through adsorption and desorption.

• Surface areas in contact with the sample stream should be minimized. Smooth surfaces will have less surface area where gas molecules can adsorb, and porous materials (such as plastics) should be avoided when building sampling systems.

• Tests have shown that sampling systems respond more quickly to changes in moisture content when the sampling flow rate is increased, since this quickens the mechanism by which molecules adsorb and desorb from the sample tube walls and other “wetted” surfaces.

• Thermodynamic conditions can also affect the physical adsorption equilibrium. Gas molecules have a higher tendency to physically adsorb at high pressures. Lowering the gas pressure in the sample system (through regulation) will tend to shift the equilibrium and desorb gas phase molecules, minimizing the resulting distortion.

• Similarly, gas molecules have a higher tendency to physically adsorb at low temperatures. Raising the temperature above ambient temperatures (through active equipment heating) will also tend to shift the equilibrium and desorb gas phase molecules. However, this is most beneficial where other sampling conditions can slow down system response times.

Components dissolved into liquids: If liquids have the potential to enter a sampling system intended only for gas analysis, filters are suggested to prevent undesirable liquids from reaching the sample line, analyzers, regulators, or any other equipment with which the gas may come into contact. Good sample system design will include methods for removal of free liquids, such as a Pitot tube return to the pipe or a drain port on a portable sampling apparatus.

In the event that liquids do enter the sampling equipment, sample systems should be designed to be thoroughly and easily cleaned. Research (Behring and Kelner, 1999) has found wet steam to be the most effective cleaning agent for removing heavy hydrocarbon liquids from sampling equipment, provided the steam does not contain treatment chemicals or corrosion inhibitors that could also contaminate the equipment. Solvents that do not leave a residue after drying, such as acetone and liquid propane, are also generally effective. Silicon grease or other lubricants that may absorb components from the sample or elute contaminants into the sample should not be used on O-rings or seals.

While sample lines and hardware in sample systems are often permanently installed and may be hard to disassemble for cleaning, sample containers are easily transportable. Furthermore, their interior surfaces come into contact with gas samples for long periods of time. For these reasons, the API MPMS, Chapter 14.1 standard requires that sample containers be purged and cleaned of liquid contaminants prior to each sample, using steam or another accepted cleaning method. Evacuating the cylinder to 1 millimeter of mercury absolute pressure or less will eliminate by vaporization any residual liquids not removed by the steam cleaning process.

Phase change due to Joule-Thomson cooling and flow dynamics can be avoided through the use of a phase diagram, such as Figure 12, to select appropriate equipment when phase change is expected. For instance, if a phase diagram of a sampling process indicates that Joule-Thomson cooling at a pressure regulator may cause sample distortion, heating of the regulator (e.g., via heat trace) is recommended to offset the cooling effect. Generally, care is recommended to provide sufficient heat to avoid condensation of gas components whenever any type of regulator is installed in a sample line or a pressure reduction occurs. This can be achieved using heated sample probes, heat trace along sample lines, catalytic heaters, and/or insulation.

As a safety margin against uncertainties in predicted hydrocarbon dew points, API MPMS, Chapter 14.1 recommends that sampling equipment be maintained at least 30ºF (17ºC) above the predicted (or calculated) hydrocarbon dew point. This margin can be reduced where documented research shows that differences between calculated and measured hydrocarbon dew points for the gas stream of interest are less than 30ºF. This guideline should be applied to all equipment that contacts the gas sample, including sample lines, regulators, filters, valves, and sample containers.

One significant change in sampling equipment has been specified by the latest editions of API MPMS, Chapter 14.1 and GPA 2166. This change involves the fill-and-empty sampling method, which is discussed in detail in both references. When samples are collected using this method, these standards require a drilled plug, a multi-turn needle valve, or another type of flow restriction to be placed at the end of a “pigtail” on the outlet side of the sample cylinder to control the sample flow rate (see Figure 14). The purpose of the flow restriction and pigtail

Page | 13

Page 14 of 15

is to move the throttling process and associated Joule-Thomson cooling far from the sample cylinder, so that condensation in the sample cylinder is avoided (George, et al., 2005). While earlier versions of GPA 2166 only allowed a drilled plug to be used as the flow restriction, other devices have been added to the method based on tests that successfully used needle valves instead of drilled plugs to shorten the time needed to obtain a sample.

Figure 14. Sampling Apparatus for the GPA Fill-and-Empty Method (adapted from GPA, 2005).

Conclusion

This paper has reviewed several chemical and physical phenomena that can change the composition of a natural gas sample, thereby corrupting the sample and leading to inaccurate analysis of the gas properties. Such phenomena can include adsorption and desorption of heavy hydrocarbons, water vapor, and other components at the walls of sampling equipment; condensation of gas mixture components due to Joule-Thomson cooling and phase change; and absorption of gas components into liquid contaminants. Research to understand these phenomena has led to techniques and standards for preventing them and avoiding their adverse effects on sample accuracy. Perhaps the most useful tool in avoiding sample distortion is the phase diagram. Proper use of the phase diagram and thermodynamic analysis of the flowing stream can allow the sample system designer to select regulators, heat trace, and other equipment that will avoid significant losses in hydrocarbon content and underestimates of water vapor content, which in turn can cause unnecessary production shut-ins or inequities in custody transfer. By explaining these phenomena to the reader, it is hoped that those who design and use sampling equipment can use the information provided herein to their benefit.

References

API, Manual of Petroleum Measurement Standards, Chapter 14 – Natural Gas Fluids Measurement, Section 1 – Collecting and Handling of Natural Gas Samples for Custody Transfer, Sixth Edition, American Petroleum Institute, Washington D.C., USA, February 2006.

ASTM D 1142, Standard Test Method for Water Vapor Content of Gaseous Fuels by Measurement of Dew-Point Temperature, American Society for Testing and Materials, West Conshohoken, Pennsylvania, 2000.

Bailar, J. C., Moeller, T., Kleinberg, J., Guss, C. O., Castellion, M. E., and Metz, C., Chemistry, Academic Press, New York, 1978.

Barajas, A. M., and George, D. L. (Southwest Research Institute), “Assessment of Sampling Systems for Monitoring Water Vapor in Natural Gas Streams,” Final Report to U.S. Minerals Management Service, Herndon, Virginia, USA, March 2006.

Behring II, K. A., and Kelner, E. (Southwest Research Institute), “Metering Research Facility Program, Natural Gas Sample Collection and Handling – Phase I,” GRI Topical Report GRI-99/0194, Gas Research Institute, Chicago, Illinois, USA, August 1999.

Page | 14

Page 15 of 15

George, D. L., Barajas, A. M., Kelner, E., and Nored, M. (Southwest Research Institute), “Metering Research Facility Program: Natural Gas Sample Collection and Handling-Phase IV,” GRI Topical Report GRI-03/0049, Gas Technology Institute, Des Plaines, Illinois, January 2005.

George, D. L., and Kelner, E., “Additions and Changes to the Latest Revision of API Chapter 14.1,” in Proceedings of the Eighty-First International School of Hydrocarbon Measurement, Oklahoma City, Oklahoma, May 2006, Class #5255.

George, D. L., and Kelner, E., “Lessons Learned from the API MPMS, Chapter 14.1 Gas Sampling Research Project”, Proceedings of the American School of Gas Measurement Technology, Houston, Texas, September 2010.

GPA Standard 2166-05, Obtaining Natural Gas Samples for Analysis by Gas Chromatography, Gas Processors Association, Tulsa, OK, USA, October 2005.

Gregg, S. J., The Surface Chemistry of Solids, Reinhold Publishing Corporation, New York, 1961.

Mayeaux, D. P., private communication, 2006.

Nored, M., and George, D. L. “A Review of the Current State and Direction of Methods for Sampling Wet Gas Flows”, Final Report to the U.S. Minerals Management Service, May 2003.

Van Wylen, G. J., Sonntag, R. E., and Borgnakke, C., Fundamentals of Classical Thermodynamics, Fourth Edition, John Wiley & Sons, New York, NY, USA, 1994.

Weast, R. C., Astle, M. J., and Beyer, W. H., editors, CRC Handbook of Chemistry and Physics, 66th Edition, CRC Press, Boca Raton, FL, USA, 1985.

Page | 15

ECONOMICS OF HYDROCARBON COMPOSITIONAL AND QUALITY DETERMINATION

The Dollars and Sense of Hydrocarbon Sampling

David Wofford, Red River Midstream, LLC

Before we get waist deep into science and technology and economics and variability in return rates based upon

the intransient effects of Federal Reserve interest rate and debt management policy on petroleum commodity

markets and related no-fault based derivative securities:

A neutron walks into a bar and orders a drink.

“What do I owe you?” the neutron asks the bartender.

“For you? No charge!”

An electron sitting at the other end of the bar jumps up with outrage and yells at the bartender ….

“Why does he drink for free and I have to pay?”

“Because you’re always so negative!” barks the bartender.

The electron turns to the attractive proton sitting next to him and asks ….

“Am I really negative?”

“Yes” said the proton.

“Are you sure?”

“I’m absolutely positive!”

So what does this little parable teach us? First of all, a bad attitude at the bar is not endearing to good service

and making friends. Secondly, chemistry jokes are only funny to a narrowly targeted audience. But more

appropriate to the topic of discussion, products are valued differently based upon their phases, uses, behaviors

and applications; so the precise sampling and measurement of hydrocarbons are critical to optimizing economic

value.

INTRODUCTION

Errors in product sampling and analysis are arguably among the greatest sources of uncertainty in hydrocarbon

measurement. The value of employing hydrocarbon sampling, conditioning and analytical processes and

procedures that help ensure the representative measurement of components is vital to the optimization of product

economic value. Subsequently, the optimization of product economic value is critical to the financial health and

performance of operating organizations. First, the fluids that we are measuring and sampling are important to

understand.

Fluids

Natural Gas

Natural gas is a naturally occurring hydrocarbon gas mixture primarily composed of methane, ethane, propane

and heavier hydrocarbons; carbon dioxide; nitrogen; and may contain other diluents or contaminants such as

hydrogen sulfide or water vapor. Natural gas is priced per the Decatherm (Dth) or Million British Thermal Units

(MMBTU). The most commonly referenced pricing index in the United States is Henry Hub.

Page | 16

Condensate & Crude Oil

Natural gas condensate is a mixture of hydrocarbon liquids that are present as gaseous components within raw

natural gas at various concentrated amounts. The heavier components condense out of the raw natural gas as

the flowing temperature of such drops below the hydrocarbon dewpoint. Some condensate is termed as

“produced condensate” as such is actually produced at the well and does not occur within pipelines that are

downstream of the production equipment. This indicates that the condensation of the hydrocarbons occurs within

the well due to the combination of pressure, temperature and fluid composition, and the produced gas emanates

from the condensate. Hydrocarbon liquids are generally categorized as condensate when such have an API

Gravity of 45o or greater.

Crude oil is petroleum liquid, generally of an API Gravity of 45o or lower, that possesses large amount of heavier

hydrocarbons. These heavy hydrocarbons are categorized as Alkanes (Paraffins), Cycloalkanes (Naphthenes),

Aromatics and Asphaltics. The amount of each class of heavy hydrocarbons in oil compounds varies greatly

across production areas.

Both natural gas condensate and oil are priced per the US Barrel (42 gallons). The pricing indexes used are

primarily classified by the geographic region of production, the most commonly referenced domestically being

West Texas Intermediate (WTI) and internationally Brent.

Operational Parameters

Initial Production (Flow) Rates (IP)

Initial Production (IP) is an important and challenging aspect of measurement system design, both from the

operating and economic perspectives. At points of production, wells will experience IP rates that are very often

much greater than the sustained production rates that will be realized as the well operates over time. This

challenges the operator to implement measurement systems that can accommodate the high initial flow rates at

the well site, and subsequently provide satisfactory accuracy as the flow rates decline. Subsequently, the

composition of the produced fluids often vary in relation to these declining rates of production as the lighter

hydrocarbons are often produced in greater amounts initially, and more of the heavier hydrocarbons are produced

in greater amounts as rates decline.

Sustained and Declining Production (Flow) Rates

As the initial production rates decline, the measurement system must be able to accommodate these changes in

flow rate and compositional variability with acceptable levels of accuracy and uncertainty. As stated previously,

this offers a challenge in designing systems that can perform satisfactorily under varying conditions. Sometimes,

a single measurement system can accommodate these changes by changing meter configurations (i.e. orifice

size). Depending upon the type of system employed, a single meter may be able to perform acceptably over very

wide ranges of flow rates. For example, an ultrasonic meter has a very broad turn-down ratio, enabling the

metering system to measure widely varying flow rates. Other situations may require that multiple meters are

employed at various flow rates in order to measure the flowing fluid accurately while accommodating the change

in flow rate. Additionally, changes in flowing conditions can result in the produced fluids having variances in

composition as the pressure and temperatures vary. Gases can contain greater amounts of heavier liquefiable

hydrocarbons as such experience changes of state within the flowing stream.

Page | 17

Multiple Sourced Fluid Flows

Another challenge to system design is when the fluid flows are from multiple sources of varying quantities and

compositions. This situation results in the fluid composition being very dynamic depending upon the quantity and

composition of the individual sources that comprise the fluid stream. As the flow rates of these individual sources

vary, so does that volume and composition of the composite stream. System design must be considerate of

these changes in order that representative compositional and quality determinations may be made as conditions

vary.

Fluid Sampling/Quality Determination Systems

It is necessary to take samples of the fluids which are being produced, measured and transported in order that the

composition and quality of such may be determined. As with volumetric measurement, product sampling and

quality determination can be performed by various means at varying costs. The justification of the means and

associated costs lie within the product variances and the subsequent quantification of content, quality and

residual values over substantive time periods.

Spot Sampling

The easiest and least expensive means of acquiring samples of fluids is per the employment of the “spot”

sampling technique. The name implies that a sample is taken that will be representative of the moment in time

and the composition of the fluid that was immediately present when the sample was acquired. The associated

costs for determining fluid composition and quality from spot sample acquisitions include required personnel time,

equipment (including transportation, sampling equipment such as cylinder and valves, shipping containers, etc.),

laboratory analysis and reporting.

Composite Sampling

A more comprehensive means of acquiring fluid samples is the employment of composite sampling systems.

Composite sampling involves the use of a mechanism that acquires a product sample periodically. The sample is

taken from the flowing fluid stream, injected and stored in a sample containment vessel for storage. Multiple

samples are collected over a time period and the stored collection of many individual samples is subsequently

subjected to laboratory analysis for determination of product composition and quality. This method of sampling

involves the additional cost of the composite sampling mechanism, as well as often employed sample

containment systems of more complex designs and capabilities. However, the composite sample reveals

information that is more representative relative to the composition and quality of the flowing fluid stream over an

extended period of time. This application is advantageous in fluid streams that are subject to compositional and

quality variances.

On-Line Sampling and Analysis

The most comprehensive means of determining fluid composition and quality is the employment of on-line

analytical systems that frequently acquire samples of the flowing fluid and transports such to a sample

conditioning and analytical system that determines the composition and quality of such. These analytical systems

are generally interfaced with other on-site measurement systems to provide a comprehensive quantification of

fluid volumetric and quality information. This information can then be immediately available to supervisory control

and business systems for the purposes of operational and commercial management in virtual real time. This

method involves the additional costs of the on-line analytical systems and associated peripheral equipment.

However, sample containment and shipping equipment, as well as subsequent laboratory analytical services are

only necessary as test or confirmation samples and analyses are desired by the operator.

On-Line systems may also be employed to measure specific components within the flowing fluid stream, such as

water vapor, hydrogen sulfide, or determine hydrocarbon dewpoint. These systems add costs to facilities, but

provide important information that enables the protection of downstream facilities that may be adversely affected

by these constituents or parameters.

Page | 18

Sample Conditioning Systems

Regardless of the hydrocarbon sampling method, sample conditioning is of critical importance and consideration

in ensuring that the sample collected is representative of the product under consideration. Why is sample

conditioning important? Hydrocarbons components will experience changes of state when subjected to

conditionings that will force the compound to reach critical parameters of pressure and temperature. This critical

parameter is the hydrocarbon dewpoint, which is the temperature, at a given pressure, at which hydrocarbons will

begin to condense from the gaseous phase to the liquid phase. Critical parameters within hydrocarbon dewpoint

determination are the cricondenbar and the cricondentherm. The cricondenbar is the maximum pressure above

which no gas can be formed regardless of the temperature. The cricondentherm is the maximum temperature

above which no liquid can be formed regardless of the pressure. A simple graphical depiction of these critical

parameters, the “Phase Envelope”, indicates these thermodynamic relationships.

i

The purpose of employing sample conditioning systems is to maintain the fluid sample within parameters that will

render a representative sample of the fluid within the system. For gas systems, this means identifying those

pressure and temperature relationships that will ensure that the collected sample is single-phase gaseous, and for

liquid samples, single-phase liquid. When samples of fluids are collected within the retrograde (liquid + gas)

region, the composition of the sample can be misrepresentative of the actual flowing product within the system.

This leads to errors in the determination of gas, liquid and recoverable NGL quantities, and directly impacts the

associated economics of production, gathering and processing.

Capital Investment and Operations

Capital investment and continuing operations costs of facilities are critical components of economic consideration.

Capital investment is necessary for project initiation and implementation. Operations requirements result in

continuous costs and must be evaluated within the context of cost/benefit. Costs of operations are attributable to

the systems and processes employed as noted previously. For example, the implementation of online analytical

and sample conditioning systems will require more capital investment than spot sampling or composite sampling

systems, but will facilitate more frequent measurement of fluid composition and calculation of critical parameters.

Generally, some of the primary categories within which these investments and continuing costs are considered

are:

1. Capital Investment

a. Exploration (Finding & Development)

b. Land and Right-of-Way

c. Permitting

d. Equipment

Page | 19

e. Construction and Installation

f. Commissioning

2. Operations and Maintenance

a. Personnel

i. Facilities

ii. Office (Information Processing, Data Storage, etc.)

b. Equipment and Supplies

i. Consumables

ii. Spare Parts

c. Compliance

i. Regulatory

ii. Safety

iii. Environmental

d. Outside Services

e. Inspection, Calibration & Repair

f. Administrative Overhead

g. Equipment Salvage Value

Ultimately, a good means of identifying and controlling operational costs is establishing such in terms of cost per

volume unit of measured product ($/MMBTU, $/BBL or $/MCFE). This enables the operator to determine the

impact such has on the comprehensive economics associated with the measured product, and realize how such

contributes to or diminishes value.

ACCURACY AND UNCERTAINTY

Needless to say, accuracy and uncertainty are of paramount importance to the effectiveness and economic

viability of measurement, sampling and conditioning, and analytical systems. Systems that cannot achieve

satisfactory levels of accuracy or minimize uncertainties will result in a costly proposition in the long run. Too

often, minimization of initial capital costs are given undue precedence over sustained accuracy and continued

cost of operations.

Conversely, disregard for capital investment employed for the sake of attempting to achieve levels of accuracy

and uncertainty that are inconsequential or without applicable merit can be ruinous to the overall viability of a

project. Acceptable levels of accuracy and uncertainty are established for many measurement parameters and

systems. These are defined within industry publications and standards by which prudent operators design,

implement and operate measurement systems and facilities. These standards should be the guide basis by

which operators formulate decisions regarding measurement and analytical system applications. To either

disregard these guiding resources, or seek to surpass such in excess, can lead to measurement systems and

associated economics that are indefensible and/or unsustainable.

A simple example may help clarify. Let’s assume that we are considering the purchase and installation of an on-

line analytical system for a facility. The facility is projected to experience average flow rates of natural gas of

10,000 MMBTUD. The “out-of-the-box” uncertainty specification for the system is +/- 0.5%. This implies that,

when properly installed, operated and maintained, we should realize accurate measurement of parameters to

within +/- 0.5% of true accuracy. Alternatively, we may be able to enhance the system per the employment of

characterization or peripheral mechanisms that will improve the system’s performance and reduce the uncertainty

to +/- 0.2%. This will have a cost that will be additive to the initial capital investment required to install the facility,

as well as increase O&M costs, and will subsequently effect the project economics. We need to determine if this

additional investment is a prudent expenditure and the addition to capital is economically justified.

Uncertainties place an “At Risk” component onto project economic evaluations. These “At Risk” components may

be considered as follows:

Flow Rate of Natural Gas 10,000 MMBTU/Day

Page | 20

(1) Measurement System Uncertainty (Out-of-the-Box) 0.5%

(2) Measurement System Uncertainty (Enhanced) 0.2%

Natural Gas Value $3.50/MMBTU

Quantity (1) “At Risk” = 10,000 MMBTU x 0.5% x $3.50 = $175/Day …. $5,250/Month …. $63,000/Year

Quantity (2) “At Risk” = 10,000 MMBTU x 0.2% x $3.50 = $70/Day …. $2,100/Month …. $25,200/Year

As can be seen, these “At Risk” values are of significance at this projected flow rate. One may conclude that the

additional capital expenditure would be justified in this case. What must also be considered in this evaluation are

the additional operations and maintenance costs. If the additional costs offset the potential in value, serious

considerations must be given to viability.

Measurement system accuracy and uncertainty are critical parameters that must be considered and managed.

As with any other parameter, such should be considered within the relevant context of the specific application in

which the system will be employed.

MEASURED PRODUCT VALUATION

Commodity Value

Another categorical aspect of hydrocarbon measurement economics is an extremely dynamic one – product

value. The impact of measurement systems on the determination, and potential misrepresentation, of the fluids

and subsequent products that such measure is of great significance. For instance, the commodity value of

natural gas itself is generally given consideration when contemplating measurement system uncertainties.

However, what about the liquefiable products resident within those quantities of gas? And, what about the natural

gas condensate liquids that may be realized within pipeline systems through retro-grade condensation? What of

the produced condensate at the wellhead? What of the produced oil? And, what about all of the things that can,

do and will happen to these products at various stages of production, gathering, processing, refinement, storage,

transportation and consumption? Suddenly, what seemed to be a simple concept of price per MMBTU is now a

multi-faceted economic review of product life cycle.

Page | 21

Natural Gas

The commodity value of natural gas itself is generally given consideration when reviewing measurement system

uncertainties. For instance, let’s assume the following conditions:

Flow Rate of Natural Gas 10,000 MMBTU/Day

Measurement System Uncertainty 0.5%

Natural Gas Value $3.50/MMBTU

Quantity “At Risk” = 10,000 MMBTU x 0.5% x $3.50 = $175/Day …. $5,250/Month …. $63,000/Year

As is evident, measurement system uncertainty is of significance in managing comprehensive economics and

value.

This issue of product valuation is of great importance when considering means and systems for the determination

of composition and quality by sampling and analytical processes. Unrepresentative samples that result in errors

in product quality determinations impact measured product economies. As an example, let’s assume that a

sample is taken of a natural gas stream, and such is contaminated with air as indicated by an abnormally high

Nitrogen content.

Correct Sample Analysis Incorrect Sample Analysis Component Mol % GPM Mol % GPM

Methane 85.000 ------ 83.501 ------

Ethane 5.500 1.476 5.403 1.449

Propane 3.000 0.829 2.947 0.814

Iso-Butane 1.500 0.492 1.474 0.484

Normal Butane 1.500 0.474 1.474 0.466

Iso-Pentane 1.000 0.367 .982 0.360

Normal Pentane 1.000 0.364 .982 0.357

Hexanes + 0.450 0.196 .442 0.192

Nitrogen 0.750 ------ 2.500 ------

Carbon Dioxide 0.300 ------ .295 ------

Total 100.00 4.198 100.00 4.124

DBTU/ft3 1239.3 ------ 1217.3 ------

Relative Density 0.712 ------ 0.717 ------

The effects of these differences in indicated composition and quality are:

Calculated volumes with correct analysis 8,070 MCF 10,000 MMBTU

Calculated volumes with incorrect analysis 8,012 MCF 9,753 MMBTU

At $3.50/MMBTU, this renders a valuation difference of $865. Furthermore, the economic effects on recoverable

NGL’s can be determined per methods and examples that follow.

Natural Gas Liquids

Natural gas liquids are extracted from natural gas at gas processing facilities, generally by processes of

absorption, adsorption, refrigeration/cryogenics and/or fractionation distillation. Each of these technical processes

has applicability to targeted product and component recoveries. As products are extracted, such will retain a

value as a recovered liquid. These values are additive economically in comprising the composite value of the

produced gas stream. The following exemplifies these comprehensive economics:

Page | 22

Component Mol % GPM NGL Price ($/Gal)

NGL Recovery Efficiency (%)

Recovered NGL GPM

Recovered NGL Value $/MCF

Methane 85.000 --------- ------------ ------------ ------------ ------------

Ethane 5.500 1.476 0.60 75.0 1.107 0.6642

Propane 3.000 0.829 0.90 95.0 0.789 0.7101

Iso-Butane 1.500 0.492 2.00 100.0 0.492 0.9840

Normal Butane 1.500 0.474 1.95 100.0 0.474 0.9243

Iso-Pentane 1.000 0.367 2.30 100.0 0.367 0.8441

Normal Pentane 1.000 0.364 2.30 100.0 0.364 0.8372

Hexanes + 0.450 0.196 2.30 100.0 0.196 0.4508

Nitrogen 0.750 ---------- ------------ ------------ ------------ ------------

Carbon Dioxide 0.300 ---------- ------------ ------------ ------------ ------------

Total 100.000 4.198 3.789

If the flow rate is 8,070 MCF & 10,000 MMBTU, the associated economics are:

Value of Natural Gas (Initial) 10,000 MMBTU x $3.50/MMBTU = $35,000

Value of Natural Gas (Post-Extraction) 7,225 MMBTU x $3.50/MMBTU = $25,287

Value of Recovered Ethane = 1.107 Gallons x 8,070 MCF x $0.60/Gallon = $5,360

Value of Recovered Propane = 0.789 Gallons x 8,070 MCF x $0.90/Gallon = $5,730

Value of Recovered Iso-Butane = 0.492 Gallons x 8,070 MCF x $2.00/Gallon = $7,940

Value of Recovered Normal Butane = 0.474 Gallons x 8,070 MCF x $1.95/Gallon = $7,459

Value of Recovered Pentanes + = 0.927 Gallons x 8,070 MCF x $2.30/Gallon = $17,206

Total Stream Value = $69,982

Difference in Value = Total Stream Value - Value of Natural Gas (Initial) = $69,982 – $35,000 = $34,982

As indicated in the example, the total value of the measured gas stream is much greater than just the price of gas

per unit of volume. The extracted value of the NGL’s greatly enhances the overall value of the measured stream.

Now, if we apply the previously used measurement system uncertainty to this new Total Stream Value:

Value “At Risk” = $69,982 x 0.5%= $350/Day …. $10,500/Month …. $126,000/Year

We realize that the “At Risk” value of the measured product stream due to measurement system uncertainty is

virtually double that of which was originally contemplated as only a valued gas stream.

Natural Gas Condensate and Oil

The value of condensate is dynamic, but often closely emulates the price of oil when stable and meeting certain

quality specifications. The economic issues related to the handling, storage and measurement of condensate are

somewhat complex, and deserve discussion in order to understand and quantify such.

As noted previously, condensate is a liquid composed of natural gas components. Most of these components are

Pentanes and heavier. However, the liquid will also contain some amounts of lighter hydrocarbons. At the point

of production, products (gas, condensate & water) are separated by mechanical means. These products are then

individually transported to points of delivery or disposal. For condensate, this often means that a dump-valve

incorporated into the condensate phase of the mechanical separation unit will open, thereby transporting

(dumping) the condensate to a tank. When this occurs, a process commonly known as “weathering” takes place.

Page | 23

As the condensate experiences the sudden change in pressure when dumped from the pressurized separation

vessel into an atmospheric tank, the fluid will begin to “roil”. This means that lighter components within the

mixture are experiencing a change of phase from liquid to gas via the change in conditions and associated

turbulence of the occurrence. As these components weather, the gases exit the tanks from the top and are either

lost to the atmosphere, or may be captured and re-injected into the gas pipeline per the use of a vapor recovery

unit (VRU). Additionally, significant amounts of the heavier components that would normally be retained within

the liquid condensate are lost as well. This occurs because of the rapid roiling of the fluid during the change in

pressure, as well as effects of ambient temperatures. All of these occurrences result in compositional and quality

changes to the fluid, and thereby affect the value of such. Mechanical stabilization processes enable many of the

heavier hydrocarbons to be retained within the liquid product rather than being flashed as vapor. The process of

controlling pressure, temperature and rate allows for the gradual extraction of the lighter hydrocarbons to occur

while retaining the heavier components as liquid product, thereby enhancing recovered volume and product

value.

Of additional economic significance regarding natural gas condensate and oil is the issue of quality specifications

associated with transportation. Most liquid pipeline tariffs will specify quality parameters and shrinkage penalties

that impact delivered volumes and valuation. Commonly, initial shrinkage of up to 2% may be applied to delivered

volumes, with additional shrinkage factors applied for deviations from certain quality thresholds. Very common

shrinkage deductions will occur at API Gravities exceeding 45o. Volumes will usually be deducted in 0.5%

increments for every 10o of excess API Gravity up to a limit, usually 2%. Trucking and rail transportation methods

will also impose quality limitations and penalties.

LOST AND UNACCOUNTED FOR QUANTITIES

Lost And Unaccounted For (LAUF) volumes of product result in significant economic impact to operating

revenues on a continuing basis. It is, therefore, no surprise that LAUF receives so much attention from operators,

industry organizations, and industry schools and seminars. The Energy Information Administration reports that in

2011, 290,000,000,000 ft3 (that’s 290 BCF for the acronym aficionado) of natural gas was lost and unaccounted

for in the United States from all sources, a 1.19% LAUF with regard to all sources of receipt and consumption. At

the average 2011 Nominal Price of $3.95 per MCF, that equates to $1,145,500,000 in lost commodity value. Of

this quantity, 200 BCF is categorized as known quantities that were vented and flared. That leaves a whopping

90 BCF of imbalance …. Unknown … Vamoose …. Whoosh …. Gone! That is 0.37% of total 2011 U.S natural

gas consumption and $355,500,000 in commodity value. ii

From the perspective of the individual point of measurement, LAUF is of relevance from many perspectives. In

gathering systems (midstream operations) LAUF is often allocated back to each point of receipt into the system

on a periodic basis, usually monthly. The amount of allocable LAUF is often capped at a maximum allowable

percentage of total system receipts. If the maximum allowable LAUF that may be allocated is 2% of system

receipts, such directly impacts the economic metrics of the producing facility. As example:

Measured quantity 10,000 MMBTU

System LAUF% 2.0%

Allocable LAUF quantity adjustment 10,000 x 2.0% = 200 MMBTU

Allocated “Lost” Value at $3.50/MMBTU 200 MMBTU x $3.50 = $700

Allocated “Lost” Natural Gas (Post Extraction) Gas & NGL Values:

Value of “Lost” Natural Gas (Post-Extraction) 145 MMBTU x $3.50/MCF = $508

Value of “Lost” Ethane = 1.107 Gallons x 161 MCF x $0.60/Gallon = $107

Value of “Lost” Propane = 0.789 Gallons x 161 MCF x $0.90/Gallon = $114

Value of “Lost” Iso-Butane = 0.492 Gallons x 161 MCF x $2.00/Gallon = $158

Page | 24

Value of “Lost” Normal Butane = 0.474 Gallons x 161 MCF x $1.95/Gallon = $149

Value of “Lost” Pentanes + = 0.927 Gallons x 161 MCF x $2.30/Gallon = $343

Total “Lost” Gas & NGL Value = $1,379

COMPREHENSIVE ECONOMIC EVALUATIONS

Now that we have touched on many of the economic facets related to the measurement, sample conditioning

and analysis of hydrocarbon products, let’s look at a more comprehensive evaluation of the life cycle of

hydrocarbons from a revenue/cost perspective. This evaluation will consider a well that is drilled and

completed, and is producing natural gas and condensate.

There are many data points (and assumptions) that must be considered in formulating this evaluation. For

the sake of this example, let’s assume the following operating and economic conditions with two distinct gas

compositions (one “representative” and one “unrepresentative” due to air contamination in the sample):

Operating Conditions

Flow Rates Gas (MCFD) 5,000

Condensate (BBLD) 500

Water (BBLD) 50

Operating Pressure (psia) 900

Operating Temperature (oF) 75

Pressure Base (psia) 14.73

Temperature Base (oF) 60

Project Life 10 Yrs

Production Decline Rates/Year 1 50%

2 20%

3 15%

4 10%

5-10 5%

Gas Processing Liquid Recovery Efficiency C2 75%

C3 95%

IC4 100%

NC4 100%

IC5 100%

NC5 100%

C6+ 100% Unit Pricing (Revenues)

Oil/Condensate $/BBL $90.00

Gas $/MMBTU $3.50

NGL’s $/Gallon C2 $0.60

C3 $0.90

IC4 $2.00

NC4 $1.95

IC5 $2.30

NC5 $2.30

C6+ $2.30

Page | 25

Unit Cost Drivers

Finding & Development (F&D) $11,000,000

Production Cost $/MCFE $1.30

Gathering & Treating Cost $/MCF $0.15

Transportation & Fractionation Cost $/Gallon $0.07

Condensate Transportation Cost $/BBL $3.00

Water Hauling & Disposal Cost $/BBL $2.00

LAUF % 1%

Consumer Price Index (CPI) 3%

Discount Rate 10% Gas Composition & Quality

Representative Sample Unrepresentative Sample

C1 84.000% C1 80.502%

C2 5.000% C2 4.792%

C3 3.000% C3 2.875%

IC4 1.400% IC4 1.342%

NC4 1.400% NC4 1.342%

IC5 1.000% IC5 0.958%

NC5 1.000% NC5 0.958%

C6+ 0.450% C6+ 0.401%

CO2 0.700% CO2 0.288%

N2 0.300% N2 4.792%

H2O 1.75% H2O 1.750%

BTU/CF 1,214 BTU/CF 1,162

Relative Density 0.708 Relative Density 0.718

The financial metrics that will most commonly be considered for project viability will be Earnings Before

Interest, Depreciation and Amortization (EBITDA), Internal Rate of Return (IRR) and Net Present Value (NPV)

with consideration of discounted cash flows. As has been discussed previously, consideration should be

given to the entire product value chain, as well as all of the capital investment and recurring operations and

maintenance costs that will be incurred. Obviously, the commodity value of the individual products is of vital

importance, but assumptions must be made as to the changes in these values over time. These changes can

be extremely dynamic and difficult to predict as such are subject to many sensitivities. For the sake of this

evaluation, all of the commodity values and recurring unit cost drivers were adjusted annually by the

Consumer Price Index (CPI) at an assumed annual rate of three percent (3%).

While it is obvious that erroneous sample characterizations would be reconciled on a timelier basis, this

example is meant to demonstrate how unrepresentative sample composition determinations can significantly

impact the economic viability of the hydrocarbon value chain. The following model depicts how these two

sample compositions would affect the comprehensive economics of this project over a ten year period.

Page | 26

In the first month alone, the economic impact is a decrease in revenue of $69,853 and a decrease in the first

year’s EBITDA of $622,035. If this trend is extrapolated over the entire project life of ten years, it results in a total

decrease in realized EBITDA of $3,233,562, a reduction in NPV of $2,199,430 and a reduction in IRR by 0.6%.

While this example represents a very simplistic variance in product compositional determination, it is readily

apparent that the economic impact of ensuring that hydrocarbon fluids are conditioned, sampled and analyzed

correctly to ensure the determination of representative compositions and quality values is of tremendous

significance.

SUMMARY

So what may we have gleaned from this venture into the world of hydrocarbon economics and the impacts that

sample conditioning, collection and analysis can have on the bottom line?

The cost of asking the questions of whether or not sample conditioning, collection and analytical systems should

be employed to ensure representative compositional and quality information……..

“No Charge”

The economic implications of misrepresenting fluid composition and quality are virtually always…….

“Negative”

And, is it economically worthwhile to take the time and effort to research and employ the best applicable means of

ensuring representative fluid compositional and quality information…….

“I’m absolutely positive”

i Image acquired from www.jmcambell.com, John M. Cambell & Co., “Variation of properties in the dense phase region; Part 2 – Natural Gas”, by Dr.

Mahmood Moshfeghian, Posted January 1, 2010. ii United States Department of Energy, Energy Information Administration, DOE/EIA -0384(2011) � September 2012, “Annual Energy Review 2011”, Page

177-194.

Page | 27

BASICS FOR NEW ENGINEERING/PROJECT MANAGERS

Brad Massey, Williams

Introduction

Many individuals challenged with the task of installing gas sampling or analytical equipment for the first time typically don’t have any background or formal training in the science of collecting, conditioning and transporting a representative sample to the end device. Often times, lacking the required knowledge to design, procure and install the equipment, the “Design Rookie” will listen and accept guidance from almost anyone that speaks with some confidence or authority on the subject. Keep in mind that the requestor was most likely given this task often times as an afterthought to a larger and much broader project. Therefore, they are instructed to get the equipment ordered and installed quickly and within budget. This approach will most usually end up either in a poorly designed and installed system, resulting in inaccurate analytical results or a system that is over designed and extremely costly. With minimal research and by following a few basic steps, the correct guidance can be found and easily followed for a correctly designed sample conditioning and transport system.

General Considerations

Have you ever heard the Stephen Covey phrase of “Start With the End in Mind”? This concept can be applied to a lot of situations but assuredly it is true in the case of designing a sample conditioning and transport system that delivers a representative hydrocarbon sample to an analytical device. If the sample system designer understands that the desired end result is to provide analytical results that represent the product as if it never left the confines of the vessel or pipe from which it was originally extracted, then this individual can begin to understand how the system needs to be designed.

Understanding the primary purpose and the reporting criteria of the analytical results is critical. Although this may not change the desired accuracy it may change where and how the sample is extracted and how the results are reported.

The system designer must also understand characteristics of the product being analyzed and how the product may be impacted from the influences of pressure, temperature, transport tubing, and sample conditioning components including filter media.

Realizing site limitations is also a critical point in design decisions. This should be a consideration as early in the project as possible to provide the project team the critical required items. There are many considerations including availability of power, electrical class locations for equipment and heat traced tubing, routing of any tubing (buried or suspended) including where and how it has to penetrate enclosures.

Installed Cost vs Ongoing Maintenance

Ongoing maintenance considerations are often overlooked especially when they have a cost component to the project installed cost. When designing a system the designer should consult with not only the analytical manufacturer but also the manufacturer of the sample conditioning equipment to get an understanding of the extent and frequency of required maintenance as well as determining consumables and their cost. If significant labor and consumable costs will impact operations budgets then the operations management should be alerted for planning and budgeting material and labor costs.

One example of this would be for an assumption to be made that on a natural gas sample system it would be acceptable to eliminate any liquid removal components at the sample extraction point and only rely on the liquid protection equipment mounted at the analyzer. This poses several problems with liquids entering and collecting in the sample line allowing phase changes to occur from temperature variances thereby ruining the integrity of the sample. This would also increase the susceptibility of the analyzer to become contaminated if a liquid blocking component failed which is intended to protect the analyzer. If liquids enter and damage valves, column sets and detectors the cost of ownership will be driven up and reliability down due to a poor engineering design.

Page | 28

Reporting Analytical Results

Analytical results from analyzers are used for considerably more than often realized. Depending on an individual’s perspective and background they may not realize how the analytical results are being used. Below are various ways the results can be used but there are likely many more not mentioned.

1. Chromatograph Analysis for Custody Volumes/Dekatherms

2. Chromatograph Analysis for Monitoring Tariff Specifications between Parties

a. BTU value

b. CO2

c. Nitrogen

d. Hydrocarbon Dew Point Determination

3. Chromatographic Analysis for monitoring individual components for plant efficiencies and other

operational needs

4. Moisture Analysis to monitor water vapor Content

5. O2 Analysis

6. H2S Analysis

7. CO2 Analysis

The understanding of the intended use is important in design because usually getting the correct answer is always the primary objective for most applications and everything else is secondary. However, sometimes getting the correct answer quickly is just as critical in many processing plant operations. If control system functions are tied to the results of an analyzer, then timely data analysis and reporting will be critical. Achieving timely information from the analyzer would require considerations in probe location, component volumes, sample pressures, tubing length and tubing diameter. Another consideration may be to utilize a sample speed loop to ensure the sample is refreshed up to the analyzer during previous sample runs.

Understanding the Product

There are many different types of hydrocarbon products found in both liquid and vapor forms within the oil and gas industry and depending on where in the system they are being analyzed plays a large part in what other factors need to be determined. These factors are critical in determining how a sample system should be designed.

For example designing a sample system immediately downstream of natural gas wellhead will often times require a great deal of consideration for dealing with high water vapor content and other contaminants found in raw unprocessed gas. On the other hand a sample system on a transmission pipeline, where water vapor content and other contaminant concentrations are consistently low, would have a completely different set of considerations.

Very rich gas streams are much more likely to have condensed hydrocarbon liquids entrained in the stream. Depending on system piping complexity in the area of sample extraction, there may or may not be an ideal point to extract the sample in order to minimize the possibility of ingesting hydrocarbon liquids. Flow restrictions caused by excessive or improperly sized transport tubing, tube fittings, and valves should be avoided to minimize pressure drops occurring in the sample transport system. These pressure drops lead to temperature reduction increasing the potential for condensation of hydrocarbon components.

Design Considerations

When designing a Sample Conditioning System there are many considerations that aren’t always apparent but are critical in the overall design. The primary reason for extracting and transporting a natural gas sample to any analyzer is to keep it representative of the flowing stream for the analysis process being executed. On the following page are general considerations that a designer should take into account.

Page | 29

Sample Conditioning System Component Design Considerations 1. Extraction Probes and Regulators

a. Single Path Probes b. Dual Flow Probes c. Membrane Tipped Probes d. Probe Regulators e. Stationary Probes f. Retractable Probes

2. Pressure Reduction a. Probe Regulators b. Multi Stage Probe Regulators c. Single Stage External Regulators d. Multi Stage External Regulators e. Heated Regulators

3. Filtration a. Membranes b. Coalescing c. Particulate d. Contaminates (ie Glycol, Methanol, Amine, etc.)

4. Liquid Removal a. Membranes b. Coalescing Filters c. Liquid Blocks d. Block and Drain

5. Transport Tubing a. Stainless Steel b. 1/8” to ¼” Diameter c. Insulated and PVC Jacketed d. Insulated, PVC Jacketed and Electrically Heat Traced

i. Meets Electrical Classification ii. Power Connections and End Terminations iii. Watts per foot iv. Length Limit for Electrical

e. Single Tube or Multi Tube f. Length Limits for Lag Time g. Speed Loops h. Buried or Overhead i. Building Entrance Transitions j. Internal Coating Availability for sulfur compounds

6. External Heat a. Applied to Regulators b. Applied to Sample Conditioning Components c. Applied to Transport Tubing d. Heat Source Type

i. Electrical ii. Catalytic iii. Steam

7. Insulation a. Covers/Boxes b. Transport Tubing c. Packaged Components d. Shelters

8. Electrical Power Availability a. AC Purchase Power b. DC Solar Power c. DC TEG Power

Page | 30

Know Your System

Regardless of who you are and where you work, there are often times when you can categorize quality aspects of system specific areas within your operating areas. This doesn’t replace sound research and sample analysis but it does provide a basis for identifying design criteria. There are many examples of this which range from specifying the wattage of heat trace tubing to specifying the type of regulator to use, single stage versus multistage regulation, in a natural gas sample system. Below is a categorization table being considered for use within an organization. Keep in mind that this is only a representative concept to examine one operator’s system.

For the purpose of this project, sample conditioning systems are defined into five different categories as outlined below. Each category identifies a general set of conditions for sample conditioning, however in every instance the operating conditions, potential for liquids and contaminants should be closely examined and designed for accordingly.

Categories

Conditions

Category 1

Transmission

Mainline &

Deliveries

Category 2

Transmission,

Gathering, or

Receipt Point

Category 3

Gathering or

Receipt Point

Category 4

Wellhead or

Gathering

BGF

Pipeline

Locations

Areas 1 & 3 Areas 2 & 4 Area 5 Area 6

Typical BTU 990 - 1050 990 - 1100 1051 -1100 > 1100

Water Vapor Consistently

<7lbs

Typically <7lbs Sometimes >7

lbs

Often to Likely

>7 lbs

HCDP Consistently

>30⁰ F above

Operating

Temp

Typically

>30⁰ F above

Operating Temp

Sometimes

<30⁰ F above

Operating Temp

Operating near

HCDP

Glycol

Presence

No Possible but not

typical

Susceptible Susceptible to

Likely

Methanol No Possible but not

typical

Susceptible Susceptible to

Likely

Other

Contaminants

No Possible but not

typical

Susceptible Susceptible to

Likely

Table 1

Categories for Sample System Designs

*In all cases full consideration of contaminants and gas composition should be considered in the design of the sample system. An equation of state software program should be utilized to determine the hydrocarbon dew point of the flowing gas stream.

In the table above it clearly distinguishes differences from various parts of a system and allows for overlap in BTU value. On the following page are some thoughts as to how you might design for each set of conditions.

Page | 31

“Category 1” Sample Conditioning System A Category 1 sample conditioning system is defined as a system that provides sample phase stability and liquid rejection at conditions where moisture or hydrocarbon liquids are not likely to be present. The occurrence of contaminants is unlikely and the gas stream is considered as “Lean Vapor Stable” with little chance of hydrocarbon liquids in the flowing gas stream or during the sample extraction, pressure reduction, and sample transport process. A “Lean Vapor Stable” gas system has a typical a BTU range of 990 to 1050 consisting of a methane concentration of 95% or greater.

A Category 1 system would provide minimal protection for the sample transport system and analyzer against the ingestion of liquids and/or other contaminants by utilizing liquid membranes or liquid blocks with filters at or near the sample extraction point. A calculation of temperature reduction caused by extreme pressure reduction should be made to insure that auxiliary heat isn’t required at the pressure reduction point. At a minimum in warm climates, transport tubing should be insulated. In cool to cold climates transport tubing shall be heat traced in a manner to maintain the temperature well above hydrocarbon dew point.

“Category 2” Sample Conditioning System

A Category 2 sample conditioning system is defined as a system that provides sample phase stability and liquid rejection at conditions where occasional moisture or hydrocarbon liquids may be present. These conditions are driven typically by a significant change in environmental factors or operating conditions that result in condensation of liquids either in the flow stream or during the sample extraction and pressure reduction process. Occasional high water vapor content is the only anticipated contaminant and/or the gas stream is considered to be “Moderately Rich Vapor Uncertainty”. A moderately rich gas stream has a BTU range of 990 to 1100 consisting of C2+ components between3% to 5%.

A Category 2 system would provide elevated protection for the sample transport system and analyzer against the ingestion of liquids and/or other contaminants by utilizing liquid membranes or liquid blocks with filters at or near the sample extraction point. Any pressure reduction equipment should be heated and insulated to prevent liquid condensation from occurring. At a minimum, in all climates, transport tubing should he heat traced in a manner to maintain the temperature well above hydrocarbon dew point.

“Category 3” Sample Conditioning System

A Category 3 sample conditioning system is defined as a system that has a higher level of anticipated hydrocarbon dew point and the likelihood of more contaminants including a higher probability of occasional high water vapor content. Contaminants that may exist are Methanol, Glycol, Amines, Compressor Oil, H2S Scavengers, Corrosion Inhibitors or Odorant.

A Category 3 system would provide elevated protection for the sample transport system and analyzer against the ingestion of liquids and/or other contaminants by utilizing liquid membranes or liquid blocks with filters at or near the sample extraction point. Any pressure reduction equipment should be heated and insulated to prevent liquid condensation from occurring. At a minimum transport tubing should he heat traced in a manner to maintain the temperature well above hydrocarbon dew point. Necessary measures should be employed to remove the effects of contaminants without interfering with the intended analytical process.

Page | 32

Category 4 Sample Conditioning System A Category 4 sample conditioning system is defined as a system that provides sample phase stability and liquid rejection at conditions where high moisture or hydrocarbon liquids are often found to be present. These conditions are driven typically by a relatively small change in environmental factors or operating conditions that result in condensation of liquids either in the flow stream or during the sample extraction and pressure reduction process. Frequent high water vapor content is anticipated contaminant and/or the gas stream is very rich. Other contaminants that may exist are Methanol, Glycol, Amines, Compressor Oil, H2S Scavengers, Corrosion Inhibitors or Odorant.

A Category 4 system would provide maximum protection for the sample transport system and analyzer against the ingestion of liquids and/or other contaminants by utilizing liquid membranes or liquid blocks with filters at or near the sample extraction point. Any pressure reduction equipment should be heated and well insulated in a suitable environment to prevent liquid condensation from occurring. Additional measures including coalescing filters and/or staged regulation should be considered. At a minimum transport tubing should he heat traced in a manner to maintain the temperature well above hydrocarbon dew point. Necessary measures should be employed to remove the effects of contaminants without interfering with the intended analytical process. A very rich gas stream has a BTU range above 1100 BTU with C2+ components greater than 5%.

Conclusion Many hydrocarbon gas and liquid streams experience many variables with very little certainty that the stream will experience stable gas quality conditions. Understanding the operator’s system composition, other quality influences and varying operating conditions is essential in designing and building a proper sample conditioning system. Understanding site conditions and limitations is also a key factor to be aware of early in the design. Asking some fundamental questions regarding the intent of the analysis reports and data outputs will help in determining how to specify and locate equipment for timely and accurate results. By identifying the location, anticipated conditions and contaminants the design can be done correctly. When correctly designed the sample should always reduce the risk of sample contamination and arrive at the analyzer in the same state and with the same properties as found within the pipe or process.

Page | 33

A CASE FOR AND AN UPDATE TO RESEARCH AND TESTING ON

THE ISSUE OF HYDROCARBON WET GAS SAMPLING

David J. Fish, Welker, Inc.

Introduction

In the last 25 years, the natural gas pipeline industry has transitioned from the supplier of clean, dry gas to the

mover of billable gas energy; clean and dry or dirty and wet. The amount of hydrocarbon product that is

transported between producer, processor, distributor and user is significant. To be able to verify the exact

composition of the product is important from an economic and product treatment standpoint. In addition, if the

best sampling procedures are followed, the potential for disputes between supplier and customer will be greatly

reduced. The importance of properly determining hydrocarbon gas composition benefits all parties involved.

The level of interest in effective and accurate gas sampling techniques is currently at a very high priority within the

natural gas industry. With current natural gas prices, exploration interests, profitability, deregulation and

consolidation of the work force, recoverable revenue must be found and reported. At large volume delivery

points, a 3-5 BTU error in energy determination can cost companies tens of thousands of dollars within a very

short time period. Accurate sampling techniques must be implemented with equal interest as that which is given

to accurate volume measurement. And, it bears repeating over and over again. Most flow equations receive the

specific gravity portion of the formula from the analysis of the gas sample. If the analysis is deprived of the total

picture, how can the volumetric answers be relied upon as representative of the total system? The error is

compounded and the integrity of the system is compromised.

Natural Gas sampling has been performed for years with techniques handed down from generation to generation.

Most of the early methods were not sufficient to meet today's requirements of accuracy and repeatability;

however, standards have been developed to reach toward these demands. The most widely known standards

are GPA-2166-05 and ISO-10715 and API 14.1 of 2006. These have already generated significant interest in

proper sampling techniques, due to a large volume of data produced during the revision work.

In the past 35 years, sampling systems have been refined to meet more exacting requirements of the industry and

sampling standards have been revised to reflect the latest reliable knowledge and techniques. The equipment

available today is truly “state of the art.” Samplers, cylinders, probe regulators, protective filtration systems,

separators, membranes, protective shut in devices for analyzers, insulated and heated enclosures and the like

are available from a number of known manufacturers.

Historically, natural gas was sampled as natural gas. Our natural gas pipelines are seeing changes within the

pipeline, relative to quality. Liquids are present for a number of reasons. One dramatic reason is simply the cost

of natural gas. Producers are trying to meet the demand and sell their product. In so doing, gas processing is

being streamlined or reduced. Liquids are being passed along in the interest of providing energy. Another reason

for the increased presence of liquids is a change in system operations. Today, we pull from our storage domes

harder and faster than ever before. This tends to increase the presence of liquids in the system. Deeper wells,

shale plays and colder pipelines in deep-water production are yet other sources of liquids. Reduced

maintenance, more flexible contracts and a host of operational considerations are in play. Hydrocarbon liquids

that are present in a natural gas pipeline have monetary value and must be accounted for. The liquids are not

being recognized in the analysis, but they are being measured as mass (density or specific gravity) by the meter.

Those liquids must be measured and analyzed in a fashion that is representative of the manner in which they

were measured as volume. MMBTU is the total of volume and energy. Sampling is the energy determination

Page | 34

delivery system for this equation, and the results have a dramatic influence on the volume measurement totals

and the bottom line profitability for the company.

Most equipment used in the gas industry is not designed to account for and handle liquids. Liquids have typically

been removed and handled as a liquid product. Today however, that is not always the case in a new multiphase

world. The quality of the pipeline product cannot be represented as accurate if the method of taking a sample

incorporates a technology or procedure that is designed to reject or isolate liquids that are present. Using a

separator, coalescing filter or a membrane designed to reject the intrusion of liquids, is not providing the complete

answer for the pipeline measurement department. If liquids never show up in the sample, but continue to be

found in headers, river crossings and drips, then something in the procedure or technology is not allowing for a

truly representative analysis to be attained. All of our sampling standards – ISO 10715, API 14.1, GPA 2166, etc.

– call for a representative sample of the flowing stream. From the Gas Processors Association publication GPA

2166-05, "The objective of the listed sampling procedures is to obtain a representative sample of the gas phase

portion of the flowing stream under investigation. Any subsequent analysis of the sample regardless of the test, is

inaccurate unless a representative sample is obtained.” And, from ISO-10715, a representative sample is, “A

sample having the same composition as the material sampled, when the latter is considered as a homogeneous

whole.” API 14.1 offers a similar statement in the latest revision, “a representative sample is compositionally

identical or as near to identical as possible, to the sample source stream”. We must capture a representative

sample regardless of the effort. Then, a technology or procedure for correctly and accurately handling the

combined sample must be developed. That is the challenge.

If we state that we have taken a representative sample from a flowing stream, then it must represent ALL of the

components present in that stream; not simply all the gas phase components of the stream. In the quest for full

knowledge of our system, we must know all of the components of the gas stream. Not all of the components act

the same, flow at the same speed or stay equally dispersed across the inside diameter of the pipeline. It is not as

simple as sampling a dry gas stream. It is noteworthy that the current and updated gas sampling standards all

make the clear statement that they are to be used in gas streams that are clean, dry, non-saturated and above

the hydrocarbon dew point of the flowing gas stream. Therefore, lessons learned in the programs need to be

presented accurately and honestly. While there are indeed few issues with taking a sample of 1012 BTU gas at

80° F, it is not likely a good practice to infer from that data that there are few issues to be seen with 1348 BTU gas

at 58° F. While it is true that a probe is not required in a laboratory test of nearly pure methane, it might be

considered questionable to extrapolate that to a 10 year old 8 inch meter run installation in North Dakota with1246

BTU gas and only a bottom tapping and a valve for a sample point. While we continue as an industry to find new

answers, we must not forget the lessons learned over the last 45 years.

Most current Gas Chromatographs boast an accuracy level of ½ of a BTU, but that should not be the comfort

zone for the measurement department. A faulty sampling method or improperly installed and maintained

equipment may alter the BTU content of the flowing stream by 25+ BTU. While the accuracy of the GC may be

considered as a given, the properly executed technique for taking the sample is certainly not a given.

On some meters we know what the possible error can be, if we know what the liquid content is. So knowing the

liquid content is becoming critical with the higher gas prices, which look like they are here to stay. Determining

the accurate dew point (leading to a liquid content answer) has been impossible with the current method of

sampling. Repelling liquids that are present in the pipeline at the sample point, is taking care of the liquids issue

for the GC….. but it is avoiding the accounting issue for the pipeline profitability and safety issue for pipeline

integrity.

Spot sampling was the primary method of acquiring a sample for analysis until the early 1970’s. This method is

still widely used today. In today's world of growing trends toward therm-measurement and therm-billing, this

method is increasingly expensive in analytical cost and man-hours, as well as a very questionable method of

assessing an accurate heating value to volume sales. It is at best a "spot" sample of what was present at the

moment the sample was taken. Minutes before and minutes after become unknown guesses. While this may be

a reasonable risk if the gas source is known by a long historical data base, most gas being consumed today is a

Page | 35

combined gas from several origins, or is switched from source to source by contractual updates; in some cases

by daily or even hourly arrangements. This author has been on location and witnessed a 62 BTU increase at a

single sample point, within a one-hour time frame. It was mainly attributed to both a substantial increase and

decrease in flow rate as well as well selection changes within the gathering grid. Also, we find typically, that the

older the well and the longer it stays in production, the higher the BTU value will become. Natural gas is an

extremely fragile product and almost every step in the production, transportation and distribution of natural gas,

will have an adverse effect on its quality. Switching wells, pressure changes, temperature changes and storage

vessels are only a few of the items that can add or subtract BTU values on the gas moving through measurement

stations. Thus, our sampling methods may not even represent the correct source in question.

The issue of sampling “wet gas” from a flowing pipeline is one that has elevated itself in importance over the past

20 years. Historically, natural gas was sampled from a natural gas pipeline that had been processed and was

relatively free of liquids. The water, condensate, or heavy hydrocarbons that could create unwanted liquids had

been basically removed by processing, separation, or filtration of some type. When a technician went to take a

sample of “dry” natural gas, he was reasonably assured that he would be able to extract a representative sample

of natural gas and return that to the laboratory for analysis. Due to the costs of production and processing and

the increased demands on the supply and numerous other considerations, that old, simple process of taking a

gas sample is no longer that simple.

This paper is presented with the desire to focus on the considerations and challenges that lie ahead. It is the

author’s desire to stimulate dialogue about the ultimate goal of measurement quality equipment for the industry as

it operates in the real world today. In order to promote discussions, there has to be an awareness of the issues

that impact the pursuit of new procedures or technology. The concerns of the natural gas measurement industry

are real and legitimate concerns. However, in the race to find answers, we should not fail to keep in mind the

lessons that we have learned and have documented to be true in the past. If we are mindful of those lessons, it

will allow us to appreciate the current advancements and accept how we arrived at them. As we look for answers

in the new world of gas pipeline measurement, we should not purge the data base of knowledge just to solve one

aspect of the problem. We must incorporate that knowledge into the pursuit of future advancements.

Admittedly, much of the basis for this paper and its content is the result of the fact that we now see gas pipelines

with more liquid content than before. Many are in fact, nearing multiphase pipelines. At times, we are led to

believe that our leaders of the past never faced or thought of this issue. This is why the author reflects in this

paper about not forgetting the lessons of the past. Here is an interesting paragraph from a paper presented at the

University of Oklahoma, to the International School of Hydrocarbon Measurement in 1982 --- 30+ years ago.

“The ability to “tame” liquids when they appear in the gas sample streams or cylinders is now at hand with

the availability of high quality new equipment. The capability of determining the heating value of the gas

at any pressure and temperature condition can be determined with reasonable accuracy by conditioning

the sample as it is directed to the measuring instrument. However, there is a need to more precisely

define a “liquid” in our contracts and state how to account for the heating value of the fluid when liquid is

present as an aerosol or otherwise. Should the Btu be determined on the gas at flowing conditions, or

should it be determined at a greatly reduced pressure and elevated temperature? Should a pressure and

temperature be selected for determining the Btu that would correspond with the average annual ground

temperature and average annual pipeline pressure? These and other points must be resolved before any

determined effort can be instituted to standardize Btu determination procedure on aerosol gasses.”

The Wet Gas Sampling matter clearly came to the forefront during the API 14.1 work. The Scope of API 14.1 was

not inclusive of Hydrocarbon Wet Gas as we know and understand it to be today. Due to the impact of this

phenomenon and reality in our operating systems today, this is an area that we must address as an industry, and

in the very near future.

Perhaps the single major issue that has created an interest in ascertaining the total picture of the natural gas

pipeline system is “wet gas.” The definition of “wet gas” as gas with more than 7 lbs. water per million cubic feet

Page | 36

is almost history. Wet gas metering is redefining how we talk about wet gas. There is a white paper written by

Dr. Parviz Mehdizadeh that describes wet gas. Wet gas, in that multiphase white paper, is defined as “gas, which

contains some liquid. The amount of liquid can vary from a small amount of water or hydrocarbon to a substantial

amount of water or hydrocarbon.” Today’s measurement issues are different from the past, but they are here to

stay. We must either return to the insistence and requirement of a clean, dry gas pipeline system (separators,

processing plants, dehydration systems, etc) or acknowledge the realities of the present. One of the biggest

challenges is the transportation or Mid-stream system operating below the HCDP. Liquids cause corrosion,

pulsation, freezing problems and basic maintenance issues that create concerns for a natural gas pipeline

system. Their presence must be addressed with an awareness of what that means to our industry.

One thing is for sure. The industry wants to push everything down the pipeline at once. It wants to sample the

entire contents at once. It wants a correct answer of total content from that single sample. Keeping some things

out of the sample is not the answer.

There was a solid attempt to work on this matter within the ISO system in TC 193 SC 3 and in 2011 WG 5 was

established and met in Nanjing, China. Then, due to United States policy with sanctioned countries and the ISO

position, that work became idle.

At the October 2012 API COPM meeting in New Orleans, the interest of Wet Gas Sampling was brought up in the

API 14.1 Working Group of COGFM. It was suggested to move this interest to CPMA. In that meeting, there was

clear interest in the matter and an Ad Hoc committee was established to present a white paper for this subject.

The people of this Ad Hoc committee are clearly aware of the need and the differences between upstream and

downstream interests in the matter. In this current interest we are not talking about water wet, but rather,

hydrocarbon wet - very rich streams that have the potential for causing real issues with current sampling

techniques and how to get a qualified representative sample from a wet gas system. These include issues of

ambient conditions and phase envelope knowledge, pressure changes that can create liquids when liquids were

not present before, and a host of similar matters.

We do not suggest that the points raised hereafter are the sole focus of the task, the answer, the correct way, or

even the appropriate wording to go forward. nor will they alone provide the answers that we seek What we do

feel is that these points can prompt some thoughts and ideas as we move forward.

Going Forward

While these issues listed below are not finalized answers or solutions, they are none the less a listing of the

issues and thoughts that will impact the pursuit of an answer to wet gas sampling.

• What definition will we use for Wet Gas?

o Dr. Parviz Mehdizadeh - “gas, which contains some liquid. The amount of liquid can vary from a

small amount of water or hydrocarbon to a substantial amount of water or hydrocarbon.”

o Lockhart-Martinelli

o API Chapter 1 Definition

o ASME MFC – 19G-2008 Definition

• Is Wet Gas – Multiphase? Is Multiphase- Wet Gas?

o For the measurement arena that downstream interests are familiar with, we are interested in

natural gas streams with entrained liquid similar to Type I and Type II as defined by Lockhart-

Martinelli, not a stream with 30 to 90% liquid. Initially, many of us are looking at a system that

the industry would typically consider being a natural gas stream with an unusually high content of

hydrocarbon liquids, when one expects (by past standards) a clean, dry supply of natural gas.

Page | 37

o Understanding that one could say yes to both, our first focus is the saturated stream we are

trying to measure and sample with conventional methods and not a multiphase production

stream, with the entire wellhead production coming by.

o After we tackle the Wet Gas issue, then we can move toward a Multiphase sampling program for

upstream applications.

• Attempt to define up stream vs. downstream keeping in mind the mid-stream folks. Is the break at the

Gas Plant?

o We have upstream, midstream and downstream interests in our industry. Perhaps our first work

will likely not focus on the production or upstream side. With the work done and the knowledge

gained in 14.1, GPA-2166 and ISO-10715, we are best positioned to work on the midstream

concerns and certainly the downstream applications. Liquids are an issue that causes

measurement errors, once you get away from production allocation. Our sampling standards are

focused on very accurate detection of BTU values, not just acknowledging the presence of

components at any level.

o Perhaps for the initial work, the plant outlet should be considered the beginning of the

downstream side, and primary area of initial attention. We fully recognize the push to move

accurate measurement closer and closer to the well head, but we need to document the impact

that such a move has on conventional sampling programs, and how to accommodate the

challenges we face when seeking a REPRESENTATIVE SAMPLE of what is in the pipeline at

flowing conditions -- all of it!! The custody transfer point has moved into the environment that we

have yet to address the physical realities for.

o Once we grasp the challenges at this level, we should have discovered useful tools and

knowledge to venture further upstream.

• What are the concerns at the well head that are different for the LDC and everyone in-between?

o The main concern at the wellhead centers on the ability to properly sample representative phase

fractions at the wellhead conditions, such that this information (phase densities and PVT

information) can be directly attributed to flow measurement (e.g. a wet gas meter) or the

application of phase behavior. Wellhead sampling becomes a necessity when representative

downhole samples are not available (or no longer applicable), or single-phase sampling and

recombination is not feasible (e.g. commingled separation facilities downstream of well).

o Are we sampling for Energy Content (heating value and Mass component % determination) or

simply for Component Identification – (it is there, but we don’t care how much is present, simply

that we know it is part of the stream). These are two distinct interests that will impact the scope

and goal of the work. Both ultimately need to be addressed.

o Operations downstream of production do not expect to deal with the potential for different types of

liquids (i.e. water, oil, emulsion, chemicals, drilling mud and lots more) you find close to the

wellhead and these fluids will create headaches for the sampling techniques and labs doing

typical natural gas analysis.

Economic Impact

What is the economic impact from poor measurement in sampling practices? While it may be initially difficult to

quantify, it can be substantial. It will depend on the application of the sample results and how they are used in the

operations of a company. Studies that have already been conducted have seen 50+ BTU swings at a single

measurement location within one hour, due to liquids, volumetric flow changes and well-head switching. That

impact is impressive when it relates to an attempt to provide accurate measurement, both component and flow

related. The impact of improper techniques today on a clean, dry system with known practices is well

Page | 38

documented and significant. Poor practices on a wet gas system is at least as significant, if not much greater.

Proper sampling techniques will also lead to improved allocation for production and pipeline operations.

Additional thoughts

• Is it better to try to separate the liquids in the pipeline and keep them from the sample apparatus, or is it

better to take a representative sample of the contents of the pipeline and then develop a methodology of

extracting the products from the cylinder? Return to gas plant separation and measurement of liquids as

liquids and gas as gas. You cannot separate the gas and liquid that easily as it comes from the line.

Some of the gas will remain with the liquid and some of the liquid will remain with the gas. Retention time

is needed to properly separate the gas from the liquid. The constant pressure cylinder is the best way to

give the gas and liquids the retention time needed to do good separation. Doing this at the line

temperature should tell you what was in the line. Naturally this will take some testing.

• If you take a representative sample of the properly mixed stream (much as we do with crude and refined

product systems), can we then (after capture) heat the contents of the cylinder until it is all vaporized, in a

controlled environment, and get a true picture of the contents of the pipeline? Or, must it be returned to

the same temperature and pressure at which it was taken.

• Of course, the fluid, liquid and gas in the pipeline must be completely mixed before the sample is taken in

order to take a representative sample. We know that we can extract a representative sample from a

flowing stream. That has been done for years. If the stream is mixed and uniform, then we can extract a

representative sample. But, can we analyze that sample correctly? What does it represent in regards to

the volume of gas and the volume of liquid to be measured? That is the question. This is where some

testing needs to be done.

Validation for the scope of this work

This issue has been driven by several concerns. Value determination has been moving closer and closer to the

well head due to allocation, royalty (land owners, State and Federal Government), joint ventures (more partners

with commercial interests), streamlining of operations, etc. Sampling has also been done at the field meter.

Sampling used to be done after the gas plant, as clean, dry pipeline quality gas – and often by the same company

or no more than two parties involved. Liquids were stripped and sampled as liquids. Now, people are concerned

with custody transfer earlier in the process. Offshore production, deeper wells, colder temperatures, richer gas

streams, shale gas formations and other considerations, all lead to higher BTU gas production and Hydrocarbon

Wet Gas streams. Our current gas sampling standards do not address gas streams that are considered as wet

gas streams or below hydrocarbon dew point (inside the phase envelope). Finally, it is agreed that the issue has

been there all along, but we did not sample at those points in the past. It was not ignored, but simply did not

come into play. Wet Gas has always been there – we just sampled after separation and processing. No need to

do it sooner, until now.

The work and scope of that work to be done

Due to the breadth of this matter, there should be participation from multiple reliable sources, including industry

companies, suppliers, government interests that have participated and followed earlier work (they understand that

this is not a 6 month project) and one or two research facilities. API 14.1 successfully utilized more than one of

the leading research bodies and it enhanced the work of the program, as well as added to the validity of the work.

There should be one leader with strong guidance, direction and input from the sampling committee, but as many

participants as are willing to work openly. The Natural Gas Industry needs to drive this due to the knowledge

present in the industry. No outside interest is knowledgeable enough to guide this project. All viable and

reasonably, potentially sound solutions should be examined and addressed in this project. Is there a better way?

Is it mechanically possible? Is it technically possible? Is it analytically possible?

Page | 39

The project will need a recognized facility with a wet gas loop. There should be at least three or four……..or

preferably a number of field locations with a variety of gas flows, to validate and field test the standard. There will

be a need for several gas labs that will participate in potential round robin testing of the samples and specialty gas

standard manufacturers for necessary standards during testing. It will also need locations that allow us the broad

range of dry to wet gas delivery points and begin to simply take samples with different technologies, and see how

they affect the integrity of the sample?

As of the writing of this paper, there is working being conducted within Pipeline Research Council International

(PRCI) on wet gas sampling. Due to limited funding they are concentrating on separating the liquids from the gas

and focusing on getting an accurate gas sample. Also, within API at this time, there is a proposed scope of work

for the solicitation of funds. This was presented by the Ad Hoc Committee on Wet Gas Sampling in API CPMA.

The proposed scope and support for the work were presented as follows:

• Proposed Project Scope:

The scope of API 14.1 and the resulting data produced from the revision of that standard, was limited to

natural gases that “are at or above their hydrocarbon dew point.” The current initial intent, while

understandably not the final scope of potential research, is to pursue the next step beyond the scope of 14.1.

Therefore, we seek to determine and quantify (in the field and laboratory) the variability of compositional and

BTU content analysis associated with gas sampling in upstream production separator outflow conditions and

similar pipeline conditions, using existing sampling methods (from API 14.1 and GPA 2166) with the target

flow regime for this initial testing program to be a Type I (as defined below) wet gas flow stream. The

Wet Gas Sampling Committee and the director of the funded research will utilize both laboratory research and

field research to study this variability issue.

Tests of sampling methods are proposed at a test facility under closely monitored conditions and in field

locations where we have a high level of confidence are within the Type I gas flow. These locations are to have

well-understood and stable process flow conditions that resemble wet-gas pipeline flows and can be

recreated in a controlled test facility.

The impact of water-wet conditions will also be followed in those locations where those conditions are

present. At this point, water-wet is not the primary focus, but is not to be ignored when present and

observable.

• Business Need for the Proposed Research:

Variability of sample results in upstream production separator outflow and Type I pipeline (transmission

pipelines for example) conditions can be related to the inefficiency of the separation process as well as the

reservoir production itself. Unprocessed gas can exhibit both water-wet and hydrocarbon-wet conditions that

will have an impact on the quality of the sample and resulting analysis. The composition and BTU content of

separated gas streams have shown wide variability, and testing should be done to determine how much of the

variability is associated with sampling and how much is associated with the well reservoir or separation

process itself. If testing shows that changing reservoir conditions or separator efficiency are not responsible

for most of the composition and BTU variability, then additional testing (Phase 2) will be warranted to

understand how the current sampling practices are rendered unreliable by the unprocessed stream

conditions. At this point, we need to address the most urgent need in our industry and focus on the area that

is most likely to bear useable fruit.

As described in API MPMS Chapter 20.3 and ASME MFC-19G, there are three ranges of interest in wet gas

measurement based on the Lockhart-Martinelli number – Type I (XLM <0.02), Type II (XLM = 0.02 to 0.3), and

Type III (XLM ≥ 0.3). Type I wet gas flows have low liquid levels, but are most likely to affect mid-stream and

upstream gas companies by causing inaccurate and unrepeatable sample compositions and BTU content,

and errors in contract payments at the metering and sampling interface. Type I wet gas flows are also

Page | 40

expected at locations where contract amounts are largest. Type II and Type III flow regimes fall outside of the

proposed scope of work, but will be considered for later research if data points to that need.

Conclusion

This is likely to be a 3-5 year project if fully funded and active, and an approximate budget of $500,000 to

$2,500,000 is anticipated. There will be a need for a very strong commitment to see the project through and an

understanding of the value of the study, as well as a need to promote this at high levels within the industry, based

on the need for such a standard and monetary impact to profitability.

The end game is to have a standard on wet gas sampling techniques with validation criteria and improved

measurement and better balances on systems.

Is there likely to be a Phase I, Phase II and Phase III approach, or not? There might be a need to present a

phased approach both from a technical viability standpoint (see what is out there, how it’s used, and limitations

first, with further steps improving on the technology) and from a timing standpoint (if our scope is too large, and it

takes too long to get something off of the ground, people will lose interest).

Participants in venues like NGSTech 2014 and other measurement forums should find this topic to be of great

interest over the next several years.

***********

"Proper Sampling of Light Hydrocarbons", O. Broussard, Oil and Gas Journal, September 1977

"Selection and Installation of Hydrocarbon Sampling Systems", D. A. Dobbs & D. J. Fish, Presented at Australian

International Oil & Gas Conference, Melbourne, Australia, 1991

"Methods, Equipment & Installation of Composite Hydrocarbon Sampling Systems", D. J. Fish, Presented at

Belgian Institute for Regulation and Automation, Brussels, Belgium, 1993

“The Importance of Discerning the Impact of New Measurement Technology”. D. J. Fish, Presented at 25th Annual

North Sea Flow Measurement Workshop, Keynote Address, Olso, Norway, 2007

Various Standards of AGA, GPA, API, ASTM and ISO

The author acknowledges the contributions made to this paper that also came from members of the API CPMA

Ad Hoc Committee on Wet Gas Sampling as we prepared a statement to API and in the preparation of the

funding request and members of PRCI wet gas sampling efforts.

djf/2013

Page | 41

1

DYNAMICS OF VAPOR-LIQUID EQUILIBRIUM IN PRESSURIZED LIQUEFIED CALIBRATION STANDARDS

Daniel Bartel, Airgas Specialty Gases

Introduction

Pressurized liquid calibration standards are use in many points of the petrochemical refining process. They are

used to evaluate incoming material, monitor process streams, assay final product purity, etc…

These liquefied calibration standards are packaged in one of three ways:

1. In a floating piston cylinder pressurized above the bubble point of the mixture.

2. In a normal gas cylinder at saturation pressure.

3. In a normal gas cylinder pressurized using a helium pressurization gas, typically at 200 psig.

Floating Piston cylinders have two chambers, separated by a floating piston. The sample gas is contained on one

side of the piston; the pressurization gas on the other side. Floating Piston cylinders are usually used when non-

condensable gases such as methane, nitrogen or carbon monoxide, are to be assayed in a liquefied gas sample.

Floating piston cylinders are the ideal cylinders for pressurized liquefied mixtures. There is no interaction

between the liquefied sample gas and the pressurization gas and the sample can be pressurized well above the

bubble point of the mixture. The disadvantages are that the cylinders themselves are expensive and limited in

internal volume. 1000cc is the most common size floating piston cylinder. Mixture weight in a 1000cc floating

piston cylinder is limited to approximately 500g to 800g.

Figure 1 - A Welker Engineering High Pressure floating piston sample cylinder

When a normal cylinder is to be used either high pressure or low pressure cylinders may be used. The mixture

will present in two phases, liquid and vapor and an Eductor or Dip Tube is required to withdraw liquid phase

sample from the cylinder. Inverting a cylinder that is not equipped with an eductor tube is in order to withdraw a

liquid sample is not recommended.

Depending on the boiling range of compounds in the mixture it can be difficult to get a high quality, single phase

sample to an analyzer when using normal cylinders at saturation pressure. This package is not recommended for

most liquefied standard applications.

When a normal cylinder is packaged with a pressurization gas, typically nitrogen or helium, the pressurization gas

is used to push the mixture out of the cylinder and supply high quality, one-phase, sample to the analyzer.

The effect of this pressurization gas on the Vapor-Liquid Phase Equilibrium (VLE) of the mixture is largely

ignored. It is assumed the pressurization gas all stays in the gas phase and that the liquid components do not

enter the gas phase. In reality this is not the case. Even with limited understanding of the VLE process would

lead us to think that some of the pressurization gas must dissolve into the liquid phase and that some of the liquid

phase compounds must vaporize and become part of the vapor phase.

Page | 42

2

The effect of this process on mixture component concentrations in the liquid phase becomes more pronounced

when mixtures are made with components of widely varying vapor pressures. Raoult’s Law tells us that the

partial pressure of each component above a solution is proportional to the pure product vapor pressure of the

component and its mole fraction in the solution.

Where:

pi - is the partial pressure of the component i in mixture

p*i - is the vapor pressure of the pure component i

xi - is the mole fraction of the component i in solution (in mixture)

Raoult’s law is only rigorously valid at infinite dilution, but it does imply that a component with a pure component

vapor pressure of 500 psia will have a higher partial pressure and, therefore a high mole fraction, in the vapor

above the mixture than one with a pure component vapor pressure of 5 psia.

A rigorous Equation of State is needed to fully understand the VLE since most calibration mixtures are neither

made at infinite volume nor with Ideal components. Unfortunately most EOS programs solve for a mixtures dew

point, bubble point or Flash. None of these solutions tells us the component concentrations in both the liquid and

vapor phases at a point not on the phase envelope.

Method

Our solution uses FlashCalcTM from Momentum Software and an Excel workbook in as iterative solver based on

the total pounds of product in the cylinder. The total pounds are then converted to molar ratios of components.

The EOS program calculates the molar partition between the liquid and vapor phases, but cannot solve for the

total moles. The actual molar content of the liquid and vapor spaces is solved by using the calculating the

amounts of liquid and vapor in the cylinder using the volume of the cylinder, calculated phase molecular weights,

liquid and gas densities and the calculated vapor phase compressibility.

The vapor mole% in the cylinder based on the actual masses of liquid and vapor in the cylinder. We compare this

volume against the Flash mole% and adjust the amount of nitrogen in the mixture until the two values agree within

0.01%

For the examples below, unless otherwise noted, initially the total mass in the cylinder was set to 20 pounds and

a pressure pad of 200psig of nitrogen was used. The temperature was 70° F. The percentage of nitrogen was

adjusted until the EOS vapor% and Real Gas %vapor% were equal and the masses of component sin the liquid

phase converted to molar percentages and recorded.

Then the amount of liquid phase in the cylinder was reduced, typically by 2 pounds, and the total gas phase

contents from the previous result were added. This sum of the cylinder contents was used as the inlet stream for

the next Flash.

Page | 43

3

Examples

The first example is a mixture of 10% propane in isobutane. I choose this example as I expected, based on the

Phase diagram of the mixture, the propane values in the liquid phase to remain constant as product is withdrawn

from the cylinder.

Figure 2

The data does not support my assumption:

Amount Dispensed

0 2 4 6 8 10 12 14 16 18

Isobutane 88.41 88.44 88.46 88.49 88.53 88.57 88.64 88.74 88.92 89.45

Propane 9.75 9.73 9.7 9.67 9.63 9.59 9.52 9.42 9.24 8.71

Helium 1.84 1.84 1.84 1.84 1.84 1.84 1.84 1.84 1.84 1.84

Table 1

The propane concentration in the liquid phase has decreased almost 5% after 12 pounds of mixture have been

withdrawn from the cylinder. Also, the liquid phase contains 1.84 mole% or helium that is diluting the liquid phase

concentration of propane. Graphic results are below in Figure 3:

Page | 44

4

Figure 3

Note how the propane concentration drops exponentially after 12 pounds have been withdrawn from the mixture.

Use of this standard, after this point, would lead to large calibration error.

I used this mixture, at the point where 12 pounds have been withdrawn, to investigate the effect of temperature on

pressurized liquid phase mixtures. The temperature used in the Flash calculations was varied from 0 °F to 130

°F. The results show that the propane concentration in the liquid phase of the pressurized mixture decreases

linearly with increasing temperature.

Degree F 0 10 20 30 40 50 60 70 80 90 100 110 120 130

%Isobutane

88.47

88.45

88.44

88.44

88.44

88.46

88.49

88.53

88.58

88.66

88.75

88.86

89.00

89.17

%Propane 9.78 9.76 9.74 9.72 9.69 9.67 9.65 9.63 9.61 9.60 9.58 9.57 9.56 9.55

%Helium 1.75 1.79 1.82 1.85 1.86 1.87 1.86 1.84 1.80 1.75 1.67 1.57 1.44 1.28

Table 2

8.60

8.80

9.00

9.20

9.40

9.60

9.80

10.00

0 2 4 6 8 10 12 14 16 18

Pro

pa

ne

Co

nce

nta

tio

n

Amount Dispensed

10% Propane/ Balance Isobutane

Page | 45

5

Figure 4

Therefore, for the most accurate results, it is important to keep liquefied calibration standards in temperature

controlled conditions.

Our next example is a similar mixture, 10% ethane – 90% n-butane. First we modeled this mixture without any

pressurization gas. Also, I was interested if decreasing the step size, the amount of product removed from the

liquid phase between calculations, would affect to results. When a mixture is used in real life the withdrawal of

product and subsequent alteration of phase equilibrium is a constant process, not stepwise as modeled here.

But, it would be impossible, using my method, to model a continual withdrawal of product based on the time

needed to iteratively solve each data point.

This mixture was modeled with both 2 and 0.5 pound withdrawal steps. The graph below shows that the error

between curves generated using the two different step sizes is almost identical. I will use a 2 pound step size in

all future estimations.

Note that in this example the ethane concentration drops below 9.5% concentration, therefore outside the certified

concentration range, after only 3 pounds of product are withdrawn from the cylinder.

-1.000%

-0.500%

0.000%

0.500%

1.000%

1.500%

2.000%

0 10 20 30 40 50 60 70 80 90 100 110 120 130

% R

ela

tiv

e C

on

cen

tra

ion

Dif

fere

nce

fro

m 7

0 F

Degree Fahrenhiet

Propane Concentration vs. Temperture

Page | 46

6

Figure 5

Next I modeled the same mixture with 200psig of helium pressurization gas. The Figure 6 shows how the

addition of pressurization helium causes a significant decrease in the liquid phase ethane concentration. This is

due to the fact that the liquid phase now contains about 1.8% helium. At this point the mixture is outside the

calibration uncertainty before any liquid is withdrawn.

Figure 6

In this case the pressure pad has an effect, but the opposite of what would be expected. This is due, mainly to

the addition of helium, which is more soluble in this mixture than the propane/isobutane mixture. Therefore, if we

7.5

8

8.5

9

9.5

10

10.5

0 2 4 6 8 10 12 14 16 18

% E

tha

ne

Co

nce

ntr

ati

on

Amount Dispensed

10% Ethane / Balance n-Butane - No Pressurization Gas

0.5 Pound Steps

2 Pound Steps

7.5

8

8.5

9

9.5

10

0 2 4 6 8 10 12 14 16 18

% E

tha

ne

Co

nce

ntr

ati

on

Amount Dispensed

10% Ethane / Balance n-Butane - With and Without Pressurization Gas

W/ 200psig He

W/O Helium

Page | 47

7

use a pressurization gas to provide high quality, one phase, liquid to our analyzer we should use the lowest

pressure to achieve the required results.

Total Sulfur, in the form of hydrogen sulfide, is a common liquefied calibration standard. This presents a different

phase system than our previous examples which have been alkanes in other alkanes. This example has a real

customer request for 100ppm of hydrogen sulfide in propane, supplied in a 5 gallon low pressure cylinder with a

dip tube and 200 psig of helium pressurization gas. The results of this experiment are shown in Table 3:

Amount Dispensed 0 2 4 6 8 10 12 14 16 18

Hydrogen Sulfide (ppm) 98.16 97.67 97.15 96.56 95.86 95.00 93.99 92.67 91.00 87.84

Propane (%) 99.11 99.11 99.11 99.11 99.11 99.11 99.11 99.11 99.11 99.11

Helium 0.885 0.885 0.885 0.885 0.885 0.885 0.885 0.885 0.884 0.884

Table 3

Graphically:

Figure 7

The concentration is within the accepted 10% certification error for this mixture until more than 16 pounds have

been withdrawn from the cylinder.

The next series of experiments all involve the same mixture:

Table 4

Component Target Conc. (%)

Methane 0.2000

Ethane 0.5000

86.00

88.00

90.00

92.00

94.00

96.00

98.00

100.00

0 2 4 6 8 10 12 14 16 18

pp

m H

yd

rog

en

Su

lfid

e C

on

cen

tra

tio

n

Amount Dispensed

100 ppm Hydrogen Sulfide in Propane w/200 psig Helium

Page | 48

8

Ethylene 0.5000

Propane 3.0

Propylene 3.0

Isobutane 25.0

n-Butane Balance

The phase diagram for this mixture shows that there is a 13.5° F separation between the bubble point and dew

point curves at 70° F.

Figure 8

This is a typical calibration mixture containing both alkanes and alkenes. When modeling the withdrawal of

product from the cylinder using one pound increments the results are similar to those previously observed. The

units used for Figure 9 are %Error from the nominal concentration.

Page | 49

9

Figure 9

Figure 10

Figure 10 shows the modeled changes in the “non-condensable” gases, methane, ethane and ethylene,

concentrations as the cylinder is depleted. Note that the methane concentration is 20% lower than the nominal

value before any product is withdrawn and then decreases linearly with the amount of product withdrawn. The

-8.00%

-7.00%

-6.00%

-5.00%

-4.00%

-3.00%

-2.00%

-1.00%

0.00%

1.00%

2.00%

-60.00%

-50.00%

-40.00%

-30.00%

-20.00%

-10.00%

0.00%

10.00%

0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00

% E

rro

r -

C3

H6

, C

3H

8,

i-C

4,

n-C

4

% E

rro

r -

CH

4,

C2

H4

, C

2H

6

Amount Dispensed

Complex Mixture w/200 psig Helium

%Methane

%Ethane

%Ethylene

%Propane

%Propylene

%Isobutane

%n-Butane

0.0000%

0.0500%

0.1000%

0.1500%

0.2000%

0.2500%

0.3000%

0.3500%

0.4000%

0.4500%

0.5000%

2468101214161820

Complex Mix - Methane, Ethane and Ethylene Concetrations vs. Amount

Dispensed

%Methane

%Ethane

%Ethylene

Page | 50

10

ethane is 5% and ethylene 7% lower than there nominal concentrations. This would be a poor calibration

standard for these three components.

The modeled concentration decay was validated for this mixture by preparing two mixtures gravimetrically and

then analyzing the component concentrations initially and after withdrawing product in two pound increments.

This work was performed at the Airgas Hydrocarbon Center of Excellence in La Porte Texas.

The correlation between the analytical results and the model were very good, well within mixture preparation and

assay errors.

Table 5

Amount Dispensed 0 2 4 6 8 10 12 14

Ave Diff

Methane 0.683

% 1.184

% 3.754

% 0.824

% 1.738

% 2.098

% 0.759

% 4.756

% 1.97%

Ethane 2.378

% 2.680

% 2.916

% 1.629

% 2.278

% 2.679

% 1.470

% 3.123

% 2.39%

Ethylene 2.943

% 3.101

% 4.190

% 2.056

% 3.961

% 2.001

% 0.647

% 2.139

% 2.63%

Propane 2.109

% 1.999

% 3.323

% 1.351

% 1.793

% 1.837

% 0.602

% 1.783

% 1.85%

Propylene 0.135

% 1.030

% 1.236

% 0.111

% 2.056

% 1.073

%

-0.032

% 0.565

% 0.77%

Isobutane 1.226

% 1.381

% 2.749

% 0.694

% 0.825

%

-0.457

% 0.420

% 1.234

% 1.01%

The results are in the next series of graphs. The line labeled (A) are the analytical result from the two mixtures,

assayed three times each and averaged. The average Standard Deviation between the two mixtures was 1.13%.

The line labeled (PR) are the results calculated using Peng-Robinson equation of state.

Page | 51

11

Figure 11

Figure 12

0.07

0.09

0.11

0.13

0.15

0.17

0.19

0 2 4 6 8 10 12 14

% M

eth

an

e C

on

cen

tra

tio

n

Amount Dispensed

Methane - Analytical vs. Estimated Concentration

Methane(A)

Methane(PR)

0.4

0.41

0.42

0.43

0.44

0.45

0.46

0.47

0.48

0.49

0.5

0 2 4 6 8 10 12 14

% E

tha

ne

Co

nce

ntr

ati

on

Amount Dispensed

Ethane - Analytical vs. Estimated Concentration

Ethane(A)

Ethane(PR)

Page | 52

12

Figure 13

Figure 14

0.35

0.37

0.39

0.41

0.43

0.45

0.47

0.49

0.51

0 2 4 6 8 10 12 14

% E

thy

len

e C

on

cen

tra

tio

n

Amount Dispensed

Ethylene - Analytical vs. Estimated Concentration

Ethylene(A)

Ethylene(PR)

2.8

2.85

2.9

2.95

3

3.05

3.1

0 2 4 6 8 10 12 14

% P

rop

an

e C

on

cen

tra

tio

n

Amount Dispensed

Propane - Analytical vs. Estimated Concentration

Propane(A)

Propane(PR)

Page | 53

13

Figure 15

2.8

2.85

2.9

2.95

3

0 2 4 6 8 10 12 14

% P

rop

yle

ne

Co

nce

ntr

ati

on

Amount Dispensed

Propylene - Analytical vs. Estimated Concentration

Propylene(A)

Propylene(PR)

24

24.2

24.4

24.6

24.8

25

25.2

25.4

25.6

25.8

26

0 2 4 6 8 10 12 14

% I

sob

uta

ne

Co

nce

ntr

ati

on

Amount Dispensed

Isobutane - Analytical vs. Estimated Concentration

Isobutane(A)

Isobutane(PR)

Page | 54

14

Conclusions

The liquid and vapor phases in a pressurized liquefied calibration cylinder do intermix. This intermixing means

that if a inert gas pressure pad is used the initial liquid phase component concentrations will be lower than

expected by the gravimetric/pre-pressure pad assay of the mixture components to the dilution of these

components with the inert pressurization. The only accurate Liquefied calibration mixtures are those prepared in

floating piston cylinders.

Shelf lives cannot be assigned to liquefied calibration mixtures made in normal cylinders due to the dynamic

effects of Vapor Liquid Equilibrium.

Using higher pressurization gas pressure does not increase mixture stability. The added amount of pressurization

gas forced into the liquid phase will decrease the accuracy of the mixture. At some point, with increasing

pressure, the mixture will change into a single phase supercritical fluid.

Mixtures prepared in normal cylinders and pressurized with helium or nitrogen may be used for peak identification

and gross component assay but should never be used to monitor critical processes or final product purity.

A+ Corporation and Airgas have been collaborating to develop a new cylinder that would use a membrane to

separate the pressurization gas from the calibration fluid. This cylinder would contain about 20 pounds of

calibration fluid, not the 1 pound that can be contained in “piston” type cylinders. The calibration fluid will not be in

contact with the pressurization gas. We hope to have this technology available in 2014.

Page | 55

1

TECHNIQUES OF COMPOSITE & SPOT GAS SAMPLING

By Royce Miller

Introduction

The most important thing in taking a sample is where and how it is taken. A sample can be taken as a spot, composite, or as a continuous sample connected to a chromatograph.

Where A Sample Should Be Taken?

A sample should be taken on the longest piece of pipeline available. Unfortunately this is usually a meter tube. Not that this is a bad place to take a sample, but swirls can be created inside a meter tube. This can cause the flowing stream to create aerosols from the liquid collecting on the walls of the pipeline. The aerosols can be picked up by the sample causing a higher BTU reading.

Should I Use A Sample Probe?

All samples should be taken through a sample probe. A sample probe is usually a valve with a piece of tubing welded to the bottom of it.

There are many different probes on the market today, including single flow probes, dual flow probes and hot tap insertion probes. Whichever probe you use, the most important thing to remember is to make sure the tip of the probe is in the center 1/3 of the pipeline. Probes that are too close to the walls of the pipe can pick up liquids. When placing a probe on a multiple meter tube station, make sure you place the probe on the tube that will be flowing all the time. Do not use a probe on a header. Headers create swirling gases and dead gas pockets; therefore, they are not a good choice for taking a sample.

Now that you have placed your probe in the pipeline you are ready to take a sample.

Composite Sampling

Composite sampling is the most representative form of taking a sample. A composite sampler takes a small bite of sample from the pipeline and injects it into a sample cylinder. If the sampler is connected to a flow computer it is possible to take samples proportional to flow. This means that if the flow rate goes up the sampling rate will increase, if the flow goes down the sample rate will decrease. The most important thing to remember when taking a sample from a flowing stream is that all sampler components must be above the Hydrocarbon dew point. If any component of your sampling system is below the Hydrocarbon dew point the heavy components of the sample will liquefy and drop out, typically causing a lean sample result. [Refer to API 14.1 for further explanation of this problem.]

Sample Cylinders

There are two types of sample cylinders on the market today, the spun-end bottle and the constant pressure cylinder.

The spun-end bottle is made of stainless steel or aluminum. The bottle usually has a ¼” NPT connection at each end for valves. If the bottle is longer than twelve inches or greater than four inches in diameter, DOT requires that a relieving device be installed on the cylinder. This is to protect the bottle from over-pressuring. When sampling into a spun-end bottle it is recommended that the bottle be placed in the vertical position. This will prevent the settling of heavy gases during the purging process. Purge the cylinder by using the procedure stated in GPA 2166-68, “Methods for Obtaining Natural Gas Samples for Analysis by Gas Chromatography.”

Page | 56

2

1. Remove plug or cap from outlet valve or cylinder.

2. Close sampler bypass valve.

3. Slowly open outlet valve and observe decrease in pressure on sample gauge.

4. Close outlet valve on sample cylinder just before pressure reaches zero.

5. Open sampler bypass valve, allowing cylinder to be filled with line pressure.

6. Repeat steps 2 - 5 the number of times shown on the table below, depending on the pressure used for purging the cylinder.

Pressure used for Number of

Purging Cylinder Purge Cycles

15-30 13

30-60 8

60-90 6

90-150 5

150-500 4

> 500 3

7. On the last purge cycle, do not fill sample cylinder again. Instead, leave it pressurized a little above zero.

8. Install plug or cap in outlet valve of cylinder.

After completing the purging process you will need to place the sampler into service.

If the BTU value is greater than 975 BTU you may want to consider using a constant pressure cylinder. If your BTU IS 975 or greater, and you are injecting into a spun-end bottle, you may be having a flashing problem. If you start off with atmospheric pressure in the cylinder and build to pipeline pressure the chain of molecules can change at different pressures; we call this flashing. When building pressure in a sample cylinder that starts at atmospheric pressure, it is possible for some of the molecules to change state (flash) as the pressure builds. Because this happens the sample may not be representative to the pipeline gas. By using a constant pressure cylinder, the gas in the cylinder, as well as the gas being sampled, will stay at pipeline pressure. This will keep the gas from flashing or changing phase.

Analyzing the Sample

After the sample has been taken it will be transported to a lab for analyzing. Remember all samples being transported under pressure must meet and comply with DOT certification. When preparing the cylinders for drawing off the representative sample there are some things to consider. When using a spun-end bottle you will need to heat the cylinder so that any hydrocarbons that might have liquefied will turn back into a vapor. When using a constant pressure cylinder you will need to apply pressure to the precharged side equal to the bottle pressure, so that when the sample is being drawn off, the pressure will remain at pipeline pressure. The other thing to consider is your lab. If you are using your own lab you are probably using a calibration standard that is close to your pipeline gas. If you are using a contract lab it is recommended that you check their standard. If the calibration standard is missing some of the properties that are present in your gas, you may not get a representative BTU of your pipeline gas. By using the procedures above you have a good chance of obtaining a representative sample of your pipeline gas.

Page | 57

1

WET GAS SAMPLING

Jay St. Amant, A+ Corporation, LLC

Introduction

Natural Gas sampling has progressed from a rote, mechanical, blind art without qualified data to support it….to a better understood application of science in the 2005 and 2006 revisions of our industry standards. With those revisions came an understanding of limitations of technology and an application of basic physics and chemistry. The same vapor-liquid equilibrium, composition, temperature and pressure relationships that have been understood in the process world for many years were applied to the sampling world. It was at that time that our industry began moving from just being able to obtain a natural gas sample that could be analyzed (one that would not harm the analyzer) …to being able to obtain analytically correct natural gas samples.

It is the goal of this paper to present an accumulation of current technology from traditional natural gas sampling together with current wet gas measurement technology that many may not be familiar with so that we can see where we stand in order to better understand where we want to go. A wise man once said that 99% of solving a problem is correctly identifying it.

Many people in our industry know of the national natural gas sampling standards, but not everyone knows the actual content of the standards. I’d like to review some of the applied science that has been integrated into our industry standards of GPA-2166-05 (1) and API 14.1 2006 (2) as it relates to wet gas sampling. Let’s start with API 14.1 2006 since the bulk of the data originated with that group and was shared with GPA 2166. Below are excerpts…actual sections of the standards.

Wet Gas Data

API 14.1 Appendix C – Lessons learned during sampling in hydrocarbon saturated and 2-phase Natural Gas streams (2)

Once the project had begun, Colorado Engineering Experiment Station, Inc. (CEESI) commissioned a wet-gas flow facility that offered the Working Group the option of performing the Field Phase under more controlled conditions, but still at operating conditions typical of many field locations. The facility offered the option of performing tests with the flowing stream at above or below the hydrocarbon dew point of the stream.

The CEESI wet-gas loop provided the ability to blend different compositions of gas in order to evaluate both the repeatability and accuracy of various sampling methods. The gas blends were generally prepared using pipeline gas, then weighing-in additional heavy components as required.

Since the project intended to evaluate methods in the wet-gas loop that were often below the hydrocarbon dew point, many of the data sets were collected under severe operating conditions where it was fully expected that achieving good repeatable analytical results was not likely. The Working Group intentionally designed the tests to penetrate the phase boundary and enter the two-phase region, resulting in entrained and free liquids sometimes being present in the flow loop.

The final scope of the project does not include recommendations for sampling in the two-phase region. The test plan defined by the Working Group for CEESI was performed under such severe conditions in the hope that recommendations could be made concerning the best methods to be used at or near the phase boundary. The data indicated that some methods might be capable of allowing sampling below the hydrocarbon dew point, but with higher uncertainties. The data also clearly demonstrated that under some severe operating conditions, when free liquids are present, none of the current methods are capable of obtaining a representative sample.

Page | 58

2

Scope of the Standards

API 14.1 1. Introduction (2)

It recognizes the critical impact of hydrocarbon dew point consideration to the overall accuracy and success of these practices and procedures.

Inaccuracies can result from using:

a. Inappropriate sampling techniques and/or equipment b. Inappropriate sample conditioning and handling c. Samples collected from non-representative locations and/or under non-representative operating conditions, or d. Inappropriate analytical methods

API 14.1 2. Purpose and Scope (2)

The purpose of this standard is to provide a comprehensive guideline for properly collecting, conditioning, and handling representative samples of natural gas at or above their hydrocarbon dew point.

The standard does not include sampling multi-phase flow (free liquid and gas) or supercritical fluids.

API 14.1 4. Definitions (2)

4.26 representative gas sample: Compositionally identical, or as near to identical as possible to the sample source stream.

GPA 2166-05 Introduction (1)

This 2005 revision of GPA (Gas Processors Association) Publication 2166 contains major changes from the 1986 version. The incorporated changes are the result of a cooperative sampling program carried out by the API (American Petroleum Institute) workgroup on Natural Gas Sampling. Data from the API project combined with data from a GPA project published in 1985 provide the impetus for this latest revision.

It is important to have a thorough knowledge of the phase behavior of the product to be sampled and of the Joule-Thomson Effect. Discussion of the Phase Diagram and the Joule-Thomson Effect can be found in API Chapter 14.1.

GPA 2166-05 1. Scope (1)

1.1 The purpose of this publication is to recommend procedures for obtaining samples from flowing natural gas streams that represent the composition of the vapor phase portion of the system being analyzed. These representative samples are subsequently transported to a laboratory and analyzed for composition and/or trace contaminants or analyzed onsite by portable or on-line chromatographs. 1.2 The methods outlined in this publication are designed for sampling natural gas from systems that are at or above the Hydrocarbon Dew Point temperature. As the temperature of the flowing stream decreases or the pressure increases to impinge upon the Hydrocarbon Dew Point, it becomes increasingly difficult to obtain a representative sample of the flowing stream. This standard does not address accounting for the liquid hydrocarbon portion of two-phase systems.

Page | 59

3

GPA 2166-05 12. Definitions (1)

12.16 Representative Gas Sample: A gas sample that is compositionally identical, or as near to identical, as possible to the sample source stream.

Vapor-Liquid equilibrium

GPA 2166-05 2. Notes (1)

2.1.3 Any component of the sampling system that separates unwanted liquids from the sample stream must be operated at flowing line temperature and pressure.

2.1.5 Ambient cooling of the sample system can cause heavy hydrocarbons to condense out of the vapor phase. The presence of any condensation in the sample inlet system will cause the sample to be non-representative. Condensation must be avoided in the entire sample system from sample probe to sample cylinder outlet or chromatograph inlet.

GPA 2166-05 8. Heating Requirements (1)

8.2.1 When using a separator or filter to remove unwanted contaminants, the separator or filter should be maintained at the same temperature and pressure as the gas in the line being sampled. Maintaining the separator at any conditions other than those of the line may cause sample distortion.

The Problem

So the scope of API 14.1 is to obtain representative samples of natural gas at or above its hydrocarbon dew point and it does not include sampling multi-phase flow (free liquid and gas) or supercritical fluids and the scope of GPA 2166 is obtaining gas samples from flowing natural gas streams that represent the composition of the vapor phase portion of the system being analyzed.

This limitation was necessary because the data from the 14.1 work done at CEESI revealed that when, free liquids are present, no technology existed to extract and transport a representative analytically correct sample of multiphase (free liquid and gas) natural gas.

Today, years later from 2005/2006, natural gas source streams have changed. We are adding Shale Gas, production gas, storage facility gas and the desire to move the analysis upstream. Now we are no longer dealing with just liquid contaminants like glycols and very small amounts of liquid in the form of occasional aerosols during an upset condition.

So what is the new “normal” amount of liquid in these applications?

How do we differentiate between the “classic” sample points and these new sample points?

How do we extract and transport a representative two-phase sample? (The age-old $64,000 question).

We must continue to use the analytically correct science learned during the work of the current versions of the industry standards to transport a liquid free or vaporized single phase sample to an analyzer or sample cylinder…following GPA 2166-05 2.1.3 that states that any separator or filter that removes unwanted liquids from the sample stream must be operated at flowing line temperature and pressure.

That science of vapor-liquid equilibrium doesn’t change.

An increase in temperature or a decrease in pressure will result in molecules vaporizing, but not all of the components will vaporize the same amount, resulting in a change to the gas phase composition.

Page | 60

4

If this occurs while sampling, we will be erroneously enriching the sample.

A decrease in temperature or an increase in pressure will result in molecules condensing, but not all of the components will condense in proportion to their concentration in the gas phase, resulting in a change to the gas phase composition. The heavies will drop out first.

This science of vapor-liquid equilibrium used to transport a liquid free or vaporized single phase sample to an analyzer or sample cylinder doesn’t change, but what about the extraction of these new two-phase samples? How do we even begin with a scope or how do we correctly identify this problem so that we can begin to solve it? We will need scientific data that shows that we have technology available that will allow the single phase limitation of our current standards to be removed.

Wet Gas Measurement

Dr. Richard Steven is currently the director of the CEESI Wet Gas Test facility. He earned his PhD in Experimental Fluid Mechanics in 2001. He has worked with two-phase flow metering projects with the U.K. government and other industry sponsors. Before joining CEESI, he worked for McCrometer (a leading global instrumentation company that specializes in the design, manufacture, and testing of flow metering technology) as their Multiphase Meter Development manager where he researched wet gas metering with differential pressure meters, and provided training in single and two phase flow metering technologies. Dr. Steven was chairman and main author of the ASME MFC Subcommittee 19 technical report on Wet gas Flow Metering. He has written several papers on the subject of wet gas and multiphase measurement.

Dr. Richard Steven, Wet Gas Measurement Class #1320.1 ISHM 2012 (4)

The subject of liquid and gas flowing together in a pipeline covers a huge spectrum of flow conditions. Within the Natural Gas production industry all combinations of natural gas, hydrocarbon liquid and water flows can be found. A common way of categorizing flows is to split the flows into two general types. That is, a gas dominant flow (with a relatively small amount of liquid) and a liquid dominant flow (with a relatively small amount of gas). The former is viewed as a wet gas, the latter is viewed as a multiphase flow.

Wet gas flow is not yet quantitatively defined by any codes and standards body. However, many, such as ASME (American Society of Mechanical Engineers) agree that wet gas flow could be defined as a Lockhart-Martinelli parameter, XLM, less than or equal to 0.3, XLM < 0.3. The Lockhart-Martinelli parameter is defined by equation 1, where mg and ml are the gas and liquid mass flow rates and ρg and ρl are the gas and liquid densities respectively. This term is rather abstract and is used mainly by wet gas meter designers and academics. It is more common for personnel involved with the front end of production to use the term “GVF” which stands for Gas Volume Fraction.

Equation (1):

XLM = √Superficial Liquid Inertia/Superficial Gas Inertia = ml/mg √ρg/ρl

Dr. Richard Steven, Wet Gas Measurement Class #8110.1 ISHM 2013 and Wet Gas Measurement Class #1320.1 ISHM 2012

Typically, the range of GVF < 80% is seen as general multiphase metering and GVF > 80% is seen, by multiphase meter manufacturers at least, as wet gas flow metering. For example,

75% GVF - 75/25 - 75% Gas / 25% Liquid = multiphase

Page | 61

5

95% GVF – 95/5 – 95% Gas / 5% Liquid = wet gas

In Equation 2 below, Qg and Ql are the gas and liquid volume flow rates respectively at actual flow conditions. Early multiphase metering designs tended to be designed for a range of GVF’s rather than the full multiphase range. A common early stipulation was the meter was to be used in the range GVF < 80%. GVF > 80% was stated to be a special case, labeled by multiphase meter manufacturers as wet gas flow, and outside the scope of the early multiphase meters. However, wet gas meter designers found it more appropriate to use the Lockhart Martinelli parameter rather than the GVF. Just as the early multiphase meter designs had uncertainties unacceptable to industry at GVF’s > 80%, wet gas meters had uncertainties unacceptable to industry at XLM > 0.3. There is no direct conversion between GVF and XLM unless the gas to liquid density ratio is stated (as can be seen from Equation 3). However, across typical hydrocarbon production pressures it became clear that a Lockhart Martinelli parameter of 0.3 was typically in the range of 85% < GVF < 95%.

Equation (2):

GVF = Qg/(Qg + Ql)

Equation (3):

GVF = √( ρl/ρg) / ( XLM + √( ρl/ρg) )

XLM > 0.3 according to API Publication 2565 - Multiphase Flow Metering and ASME Wet Gas Flowmetering Guideline ASME MFC-19G would be a Type III wet gas. Type II being in the range of XLM = 0.02 to 0.3 and Type I being XLM < 0.02. It has been suggested that our first step in sampling Natural Gas that is not at or above its hydrocarbon dew point begin with Type I wet gas, XLM < 0.02.

Again with no direct conversion between GVF and XLM, for the majority of industrial applications, Type I wet gas corresponds to a typical range of GVF > 99.8% (less than 0.2% liquid).

Philip Lawrence published a paper at the Wet Gas Flow Measurement Workshop in Rio De Janeiro, Brazil December 2011, Wet Gas Sampling in the Upstream Area. (5)

In this paper, he states:

Due to the nature of the problem (flow homogeneity) many wet gas sampling systems may only be limited to High Gas Volume Fractions (GVF) values, the data presented in this paper is limited to a GVF of around 99% and pressures about 90 bar g. (5)

The flow regime chart below shows the expected region of operation indicated by high superficial gas velocities (red elliptical highlight). (5)

Page | 62

6

The problem of flow homogeneity discussed by Lawrence is also noted by Dr. Steven

Dr. Richard Steven Paper Multiphase Flow Measurement Class #8110.1 – 2013 (3)

Wet gas being the result of inefficient multiphase separators due to the push to maximize financial return using existing equipment, or aging main well dynamics that change, or additional satellite wells added.

Sampling natural gas with expected entrained liquids – multiphase – wet gas

Velocity difference between phases is called “slip” and the ratio of phase velocities is called the “slip ratio”. In homogenous flow the slip ratio is unity. This is an extremely rare flow pattern in actual hydrocarbon production flows and it only exists at extremely high pressure and flow rates. Unfortunately, as the flow pattern is seldom homogenous, the multiphase meter manufacturers have to account for this by taking account of the slip ratio with correlations based on “slip models”.

Equation (4):

Slip Ratio = K = Ug/Ul (where Ug = Average Gas Velocity and Ul = Average Liquid Velocity).

Page | 63

7

There is often confusion between the meaning of Gas Volume Fraction (GVF) and Gas Void Fraction (αg).

Equation (2) above:

Gas Volume Fraction (GVF) = Qg/(Qg + Ql)

Equation (5):

Gas Volume Fraction (GVF) = AgUg/(AgUg + AlUl) = Ag/(Ag + (Ul/Ug)Al) = Ag/(Ag + (1/K)Al)

Equation (6):

Gas Void Fraction (αg ) = Vg/(Vg + Vl) = AgL/(AgL + AlL) = Ag/(Ag + Al)

Where: Ul = Average Liquid Velocity

Ug = Average Gas Velocity

Vl = Liquid Volume in Pipe Section

Vg = Gas Volume in Pipe Section

Al = Cross Sectional Area of Liquid Flow

Ag = Cross Sectional Area of Gas Core

A = Cross Sectional Area of Pipe

L = Unit Length

Equation 5 shows the definition of GVF in terms of cross sectional flow areas and the slip ratio. It can be seen that the Gas Volume Fraction and Gas Void Fraction are not the same unless the slip ratio, K, is unity. Therefore as slip exists between phases in most real hydrocarbon production multiphase, two-phase and wet gas flow applications, the Gas Volume Fraction and Gas Void Fraction are in the vast majority of cases not the same parameter.

Entrained Liquid Dispersion

Dr. Richard Steven, CEESI; Gordon Stobie, ConocoPhillips; Andrew Hall, BP; Bill Priddy, BP

Horizontally Installed Orifice Plate Meter Response to Wet Gas Flows Appalachian Gas Measurement Short Course 2012 (6)

The liquid phase dispersion is dependent on the balance of forces acting upon it. The liquid in a horizontal wet gas flow is driven downstream by the gas flow. As the line pressure and/or the gas flow rate increase the energy available to drive the liquid increases.

Therefore, the entrained liquid dispersion in a wet gas sample could be an ever changing dynamic condition.

Some of the current technology for sampling wet gas

1. Petrotech IsoSplit Separator from Sampling Expo Fluids Expogroup.com Info Sheet M1003/rev_12/12

ISO-Split isokinetic sampling technique:

Page | 64

8

• Isokinetic sampling thru an injectable probe positioned in the gas outlet of the separator

• Analysis of upstream & downstream samples taken with the probe

• Determination of separator efficiency and measured ratio for recombination of gas and liquid samples

2. Jiskoot/Cameron Mass Phase Wet Gas Sampling System for very high GVF Wet Gas sampling is comprised of take-off probe positioned immediately downstream of a homogenized sample and a bypass loop in an enclosure, containing a mass flow meter and molecular sieve and a sample container. The wet gas flows thru a molecular sieve, which traps the liquids from the gas stream. The liquid is retained in the sieve and the mass flow of the dry gas is measured by the flow meter. A sample of dry gas is also taken for analysis. At the end of the sampling process the increase in mass of the molecular sieves gives an accurate measurement of the mass of liquid removed for the gas. The liquid mass is established by comparing the mass of liquids collected with the dry mass of the gas from which it was extracted. This ratio can be used in a variety of equations to directly compensate the wet gas measurement totals. Once removed from the system the sample of dry gas can be analyzed in the laboratory. The liquid trapped in the molecular sieve can be removed using a laboratory sample extraction method and the liquid samples analyzed.

Some of the current technology employs isokinetic sampling which requires taking a sample under such conditions that there is no change in momentum, so that the sample will be representative of the gas and liquid in their respective proportions, the exact same velocity as the sample in the pipeline and the practical reality of that can only hope to be realized if you have a homogenous sample – no slip.

Is it practical or even possible to simulate and test in controlled conditions the real, typical, multicomponent Type I Wet Gas samples in the liquid forms that actually exist in the field?

Dr. Richard Steven Paper Multiphase Flow Measurement Class #8110.1 – 2013 (3)

The requirement for multiphase meters is wide spread throughout the industry. The range of field conditions is extremely wide and whereas the available multiphase and wet gas facilities offer significant ranges of flow conditions, even between them, they could never seriously hope to cover the whole range of the hydrocarbon production flow conditions.

Referring back to API 14.1 Appendix C (2) testing that was done with CEESI...where the project intended to evaluate methods in the wet-gas loop that were often below the hydrocarbon dew point. Many of the data sets were collected under severe operating conditions where it was fully expected that achieving good repeatable analytical results was not likely. The Working Group intentionally designed the tests to penetrate the phase boundary and enter the two-phase region, resulting in entrained and free liquids sometimes being present in the flow loop.

We can see that one of the problems is the ability to quantify the amount of liquid at the probe in an environment where the liquid can change phase. Liquid may be injected, but vaporization of the liquid is almost certain. Also, the amount of liquid required to obtain certain GVFs cannot come from the gas alone, at least in certain set volumes. With that being said, we may basically be left with only two choices in the CEESI loop referenced in Appendix C of API 14.1 as listed on the following webpage:

http://www.ceesi.com/WetGasMultiphase/WetGasMultiphaseFacility.aspx

1) CEESI Wet Gas Loop controlled liquid – EXXSOL D80

Page | 65

9

Exxsol D80 Fluid is characterized as “de-aromatized” (non-aromatic chain) hydrocarbon solvent. The major components are normal paraffins, iosparaffins and cycloparaffins. The product contains very low levels of aromatic hydrocarbons.

Chemical Name: Hydrocarbons, C11-14, n-alkanes, isoalkanes, <2% aromatics

C11-C14 are not typical nor are they ideal liquids to be able to sample and analyze.

2) Decane – C10 – a fluid that has a history of remaining in the liquid state in the CEESI flow loop. Again, C10 is not the ideal liquid to sample and analyze (vaporize and analyze).

It may be difficult or impractical to simulate and test in controlled conditions the real, typical, multicomponent Type I Wet Gas samples in the liquid forms that actually exist in the field. More Questions than Answers Can current membrane tipped probe technology that complies with the current standards and is designed to be used inside the pipeline at the flowing temperature and pressure vapor-liquid equilibrium be used if the sample is truly homogenous for Type I Wet Gas? If it were possible to simulate the multicomponent Type I Wet Gas, would a homogenizing means be possible for all types of liquids and combinations of liquids?

What happens to the homogenizing means when you have different hydrocarbon liquids, or the addition of water, glycols, methanol, scavengers, etc.?

Do different compositions, different surface tensions, different viscosities, different velocities, different piping configurations require different mixing techniques to have a homogenous mixture?

Does the attempt to mix and homogenize the sample actually change or distort the original conditions?

How would you know that the homogenizing means, if initially achievable, is reliable and efficient at all times? When pipeline conditions change? …at all changing velocities and flow rates?

Many different dynamic changing pipeline conditions make the homogenous mixing requirement very difficult.

Is true homogenous mixing possible in every location and for every application?

If it is not, would velocity slip have to be considered? If so, would isokinetic sampling even be possible with velocity slip equal to any value except unity? Conclusion Hopefully this paper has sparked even more questions that need to be asked and the need for discussions to take place as we as an industry begin to try to solve this problem of wet gas sampling by first correctly identifying it. Even the definition of what wet gas is has not yet been resolved and accepted. The process of correctly identifying all the requirements and expectations of an analytically correct wet gas sample will require a significant amount of time and resources, and we must have qualified data (applied science) to support any standards or recommendations that we make. We must be careful to not just settle for a wet gas sample that can be analyzed (one that will not harm the analyzer) but instead we must continue to work together until we can extract and transport reliable and repeatable analytically correct samples of wet gas.

Page | 66

10

References:

1. GPA-2166-05 Obtaining Natural Gas Samples for Analysis by Gas Chromatography, Gas Processors Association.

2. API 14.1 2006 Manual of Petroleum Measurement Standards Chapter 14 – Natural Gas Fluids Measurement Section 1 – Collecting and Handling of Natural Gas Samples for Custody Transfer Sixth Edition, February 2006.

3. Dr. Richard Steven – CEESI Wet Gas Test Facility Director – Multiphase Flow Measurement ISHM Class #8110.1 – 2013.

4. Dr. Richard Steven – CEESI Wet Gas Test Facility Director – Wet Gas Measurement – ISHM Class #1320.1 – 2012.

5. Philip Lawrence - Cameron V&M Director of Business Development - Wet Gas Sampling in the Upstream Area - Wet Gas Flow Measurement Workshop in Rio De Janeiro, Brazil December 2011.

6. Dr. Richard Steven, CEESI; Gordon Stobie, ConocoPhillips; Andrew Hall, BP; Bill Priddy, BP

Horizontally Installed Orifice Plate Meter Response to Wet Gas Flows - Appalachian Gas Measurement Short Course 2012.

Page | 67

YESTERDAY IS GONE

Gary W. Hines, Southern Gas Association

Introduction

“She accepted.” As I sat and listened to those words, I had a small sense of surprise. The actual acceptance of the job at the Southern Gas Association was not what surprised me. I felt confident we had hired the right person and felt we had been genuine about describing our culture and the actual job requirements. No, what surprised me was how we found the applicant. I knew that LinkedIn® was a toolset to allow business professionals to meet one another and to network with others through peer networks. I knew that there were people that moved in and out of jobs and careers in the LinkedIn world and that people would rely on their developed network to hopefully find their next job. No – what surprised me was that I, a Generation X’er, used the tool to assist in finding one of our newest employees.

Background

I have now worked at the Southern Gas Association for about three years. During those three years, I have come to appreciate our industry even more than my previous 15 years in the industry. One area I have become increasingly inquisitive of is the differences in generations all the way from those that have retirement in their crosshairs to those that can’t even spell retirement and have the ink still fresh on their college diploma. I myself, as previously mentioned, am a Generation X (more on what this means later). I have worked with a variety of generations and through my current work have met all generations in the workforce. In addition, I also have three young kids that will most likely be of a different generation than those that are currently entering the workforce. I have come to appreciate, as well as acquire headaches from, the differences of these generations in the workplace. We’ll briefly cover these different generations, what makes each different, and I’ll provide a few personal insights on some things to consider when working with different generations.

The Generations

Most have probably heard of most or all of the generations covered here, so that information will not likely be new. However, I think it is useful to provide a review of those key differentiators to better understand why some generations seem to claw at each other. Something to keep in mind is that different researchers break the generations at different years. This is not to say that the information included herein is correct and other information is wrong. It just means you as the reader should be aware of this when considering the breakdown. Also, if your birth year is within a couple of years either way of the generation break, you can actually probably claim traits of both generations on each side you fall. As an example, “Baby Boomer” for the purpose of this review ends at 1964 with the “Generation X” picking up at 1964. Therefore, if your birth year is 1966 – you may still have traits of a Baby Boomer even though technically by the dates given here you fall within Generation X. Don’t fret. Use this information as a guide and not as an absolute.

Baby Boomer

Simply put, this is the era post World War II from about the years of 1946 to 1964 with about 76 million people. Figure 1 shows the rise in the birth rate for this generation and how it dramatically tapered off in the early 1960s. Due to the rise of the economy following World War II, this generation often (relative it its predecessor generation) was much more affluent than those before them. Although others obviously came before them, the Baby Boomer generation was the first generation that really became defined due to the rise of personal marketing by firms looking to sell their goods to a growing population. Memorable events during these times include the Cuban Missile Crisis, the assassination of JFK, and man walking on the moon. An interesting note shows that by 2020 that about

Figure 1: US Birth Rate (per 1000 population)

Page | 68

25% of the workforce will be over the age of 55.1 This is shocking and further demonstrates the need to ensure that we reduce the “brain drain” as this generation moves on to the next stage of life – retirement. But why is this generation staying in the workforce? Two reasons that come to mind quickly are that people are just living longer today than they did 30-50 years ago with medical advancements and secondly, an economic rollercoaster between 2007-2009 that effectively robbed many that were ready for retirement of their nest egg – and now they find themselves somewhat starting over.

Generation X

Also known as “Gen X,” this generation most commonly has book ends of 1964 to the early 1980s. Given this title as a known change from the Baby Boomer’s it was originally titled as the generation “X” – and the term stuck. Including between 82-84 million people (the number varies depending on which years you include), this generation was the first to be known to have a volunteer spirit with almost 30 percent donating their time to others.2 One definition given to this generation is the “MTV Generation” due to the creation of music videos and music styles such as heavy metal, punk rock, grunge, and hip hop. The U.S. Census Bureau in their 2009 Statistical Abstract also reported that, “Generation X statistically holds the highest education levels when looking at current age groups.” Lastly, there also seems to be a re-sparked interest in entrepreneurial spirit among Gen X’ers with a renewed economy of the 1990s through the early 2000s.

Millennials

This generation falls most commonly between the early 1980s and the early 2000s. Again, specific years for the end of this generation vary depending on which researcher you read but regardless they all seem to agree that, per Figure 1, a resurgence of birth rates prevailed between these years (also giving it the title as the “Echo Boomers” when compared to birth rates during the Baby Boomer generation). As more and more research is performed for generational gaps, this generation seems to be punished with the most titles of any other previous generation. Titles such as Generation We, Global Generation, Generation Next, Net Generation3, and then just simply Generation Y as a follow on to Generation X. Here is something interesting about the cutoff birth years of each generation. This is mostly not known until children hit the age of about 11-13. Why? Marketing firms only then start figuring out the real traits that separate them from their older siblings to better know how to entice them and sell to them. This is why we are now learning a new differentiation between them and the next generation, Generation Z (Of course!). Culturally, there are some unique socioeconomic impacts that this generation experienced as children – most notably the Columbine (Colorado) school shooting and being the first generation to grow up with the internet from an early age. One author, Ron Alsop, described this generation as the “Trophy Generation” where all (well, most) kids competed in some form or another of team and competitive sports and all received a trophy whether they win or lose.4

Generation Z

Also called iGeneration, Gen Tech, Gen Wii, Net Gen, Digital Natives, Gen Next, Post Gen5 this generation is where we currently know the least, but researchers are certain that a differentiation from the Millennials began in the early 2000s. While most in this generation were too young to know the impact, they have grown up with September 11, 2001 forever being etched in their minds. Some researchers actually think this will drive this generation to be more “homebodies” than before due to their parents not wanting to lose them or let them go. Truthfully, there is a lot yet to learn about this generation – including a good name. More will be defined over the coming 5-7 years that will further shape this generation’s differences from its predecessors. Regardless, the key is to know that a new generation is growing up in this country that will be different that those that have gone

1 Chosewood, L. Casey (July 19, 2012). "Safer and Healthier at Any Age: Strategies for an Aging Workforce". National

Institute for Occupational Safety and Health. Retrieved 2014-01-10. 2 "Volunteering and Civic Life in America: Generation X Volunteer Rates". Corporation for National and Community

Service. November 27, 2012. Retrieved 2014-01-10. 3 Shapira, Ian (July 6, 2008). "What Comes Next After Generation X?". The Washington Post. pp. C01. Retrieved 2014-01-

10. 4 Alsop, Ron (October 13, 2008). The Trophy Kids Grow Up: How the Millennial Generation is Shaking Up the Workplace.

Jossey-Bass. ISBN 978-0-470-22954-5. 5 Horovitz, Bruce (May 4, 2012). "After Gen X, Millennials, what should next generation be?". USA Today. Retrieved 2014-

01-10.

Page | 69

before them – and they must also find ways to work with those generations that are younger than them and most likely not yet born.

So What!?

So what does all of this mean? Why should I care? No matter if you are reading this and are like me and can barely spell retirement or you are planning your retirement party in 2014 – it matters. While technology for instance changes how we do things in our industry – all must know the basics of gas quality and how it impacts the transportation of natural gas. I doubt there are arguments with that, but how different generations learn those basics are enough to make us pull our hair out. Some learn by reading a manual. Some learn by watching a tutorial video. Some learn by simply performing a task. Most learn by some combination of all of these efforts plus others. Some of the following may seem over-simplified, but there are times when simple is best and starting with some basic needs may just be the answer needed to solve a challenge such as learning how to work with and communicate with multiple generations in the workforce. What you will not find in these antidotes is a “finger point” from one generation to the other. I am guilty as well of telling myself that the younger generation(s) need to stop and listen and learn from me. However, this approach does not level the playing field. I must be willing to share but I must also be open to learning something from others as well – younger and older. With that, here are some considerations when trying to work with those varying traits and how we as people learn.

Evaluate Your Perspective

I remember several years ago being in a leadership workshop and the instructor asked the question, “Do those in Europe drive on the wrong side of the road?” My immediate answer, in my head of course, was a resounding “YES!” He gave little room for debate or answers and gave a resounding “NO!” Well, I knew he was wrong, but wanted to hear his justification…I was ready to provide rebuttal. To provide support for his answer, he simply asked another question of, “Why is driving on the right side of the road the ‘correct’ side?” Well, I obviously had no answer for this – and had just lost the argument. As a different analogy – I offer a look at Figure 2.Take

a good look at it and then honestly answer how long it took you to recognize what country this map represented. Just because it is a different perspective does not make it wrong. However, seeing it from someone else’s perspective is not always easy – it just requires patience and effort.

Ask Questions

Yeah – it’s that simple. Most often when there is a misunderstanding or a frustration it is out of lack of the full picture from the other party. While not a television viewer myself – I do get an opportunity to listen to radio and with satellite radio, I actually enjoy listening to Bloomberg Radio with Pimm Fox. Specifically, he is challenged during his afternoon show to determine who his mystery guest is with only a few clues and then having the opportunity to ask questions.6 Pimm is a smart guy and he often beats the 90 second timer by just simply asking the right questions…and sometimes asking a lot of them.

Expand Your Generational Knowledge

This does not mean you have to go to your local library and check out six books dealing with generational gaps. There is certainly nothing wrong with that, but in reality, I find that there are multiple articles posted through LinkedIn and other social media channels that will allow all of us to better work with peers all the way from the Baby Boomers to the Millennials. As an example, an October 2013 article posted on the website Revenade indicated that there was a large discrepancy between where new college graduates believe they are and where

6 http://www.youtube.com/watch?v=lbv6r2tvK4A, Serendipity3channel. “Stephen Bruce, mystery guest on Bloomberg's

Taking Stock with Pimm Fox” YouTube. September 9, 2012. Web. Retrieved 2014-01-12

Figure 2: Google Earth Maps

Page | 70

the hiring managers believe they are when it comes to their organizational skills.7 What did the article say about those skills and who was “right” and who was “wrong”? Well, you’ll have to read the article yourself – but let’s just say that there is another twist to how new graduates are or are not being prepared for the workforce. A blog entry was also found from Jessica Lawlor that provided some perspective from a new college graduate and how she challenged herself to enter the workforce.8 In a few minutes, the reader picks up eight tidbits from her perspective on what she has done to better prepare for today’s workforce. Interestingly, her last point is Have Fun! I couldn’t agree more.

Be Yourself

Whatever you do, you need to be yourself and not mold or conform to some predetermined set of standards that someone else feels you should adhere to. Diversity in many different ways is what makes our workforces successful. Even companies such as Google, Microsoft, Oracle, and other high-tech companies have diversity within their walls. Yes, they may be heavily weighted to one generation or the other, but they still have diversity. They, like others, are better for that. Now this does not give a “hall pass” to us to be stubborn and only follow our way and not learn from others, but it should allow us to include our individual approach to our work. You play an important role in your organization. It reminds me of a commercial that Honda Automotive developed for their European car line. There is a lot of reading available for the time it took for Honda to get the final cut of the two minute commercial9, but in the end each part (literal car parts) plays a critical role in providing the end result. Honda’s only words in the commercial are, “Isn’t it nice when things just…work?”10 Yes it is. And it takes everyone doing their part to make it work.

Closing

Again, I don’t know that there is anything really earth shattering in this paper. It’s really a matter of recognizing what each person can offer and then taking advantage of that. We all should certainly be open to learning from one another. Frankly, I find it exciting to learn from generations on each side of me. It does somewhat scare me to know that my oldest child will be entering the workplace in another 5 years. What will that look like and what will that mean for her generation? I don’t yet know as I’m still trying to figure out her generation where they exist today. But again, that’s the excitement of the journey. Don’t think that the journey is only for others. It’s for everyone. Don’t wait on someone to tell you when to start that journey if you have not yet started learning from others and what they can offer and what you can offer them. Start today.

7 Drott, Ellen (October 30, 2013). “Are Recent College Graduates and Hiring Managers on the Same Page?” Revenade.

Retrieved 2014-01-12 8 Lawlor, Jessica, “Reflections from the Real World: Advice for New College Grads.” Weblog entry. Jessica Lawlor – Life

begins at the end of your comfort zone. May 13, 2013. Retrieved 2014-01-12. 9 http://www.youtube.com/watch?v=NOY4JThl1Bs, tom dissington. “Making of the Honda Cog” YouTube. December 4,

2008. Retrieved 2014-01-12 10 http://www.youtube.com/watch?v=Iyh19A6CmBw, Bayridgehonda. “Honda The Cog HD” YouTube. January 30, 2013.

Web. Retrieved 2014-01-12

Page | 71

CNG SAMPLING Darin L. George, Ph.D., Southwest Research Institute

Introduction

Recent advances in natural gas production methods, particularly hydraulic fracturing, have led to increased gas reserves and lower natural gas prices in the United States. These factors have encouraged the use of compressed natural gas (CNG) as an alternative fuel for commercial vehicles and light-duty personal vehicles. Natural gas is expected to be the fastest-growing fuel in the transportation sector, with an expected average annual growth rate of almost 12% per year between now and the year 2040. Natural gas vehicles (NGVs) are attractive because of the low price of CNG on an energy-equivalent basis relative to gasoline, and because of the lower greenhouse gas emissions of CNG compared to gasoline and diesel fuels. In several cities, passenger NGVs have been granted access to HOV lanes to encourage their use as personal vehicles.

There are several challenges to the widespread acceptance of NGVs, however. An economic challenge is the higher initial cost of an NGV over a gasoline vehicle. Convenience is another concern, particularly because of the need for more CNG refueling stations in the U.S. A key technical challenge is maintaining the quality of the CNG delivered at refueling stations at the level required by NGV manufacturers. While natural gas transmission pipelines impose gas quality requirements in their tariffs, natural gas can deviate from these requirements before it is dispensed into the vehicle as CNG. Inferior gas quality can result in various problems for CNG applications, from shortened NGV engine life and degraded fuel economy to operational problems at the refueling stations themselves. Thus, it is critically important to have accurate and timely data on CNG quality, particularly on factors that can affect NGV performance. Universally-accepted methods of sampling and analyzing CNG are needed to help station operators and NGV manufacturers monitor CNG quality and avoid these problems.

This paper reviews the potential gas quality concerns facing NGV users and discusses how the natural gas industry might address these issues. The paper begins by discussing the recent introduction of NGVs to the U.S. market and the operation of NGV refueling stations. Key CNG quality parameters are then described along with their effects on NGV performance. The paper ends with some thoughts on how existing natural gas sampling and analysis methods could be modified to help the CNG industry avoid fueling problems.

Natural Gas Vehicles

It is estimated that about 123,000 NGVs are on U.S. roads today. Most of these vehicles are bi-fuel trucks and commercial vehicles that can run on either CNG or a second fuel, with separate tanks and fuel lines for both types of fuel. For example, Ford has adapted their F-150 light-duty truck to run on both liquefied petroleum gas (LPG) and CNG. Chrysler markets an NGV version of the Ram 2500 Heavy Duty pickup that uses CNG as the primary fuel and switches to gasoline when its CNG tanks are empty. GMC also sells bi-fuel commercial-duty pickups (the Chevy Silverado 2500 HD and the GMC Sierra) and bi-fuel cargo vans. The estimated ranges of these vehicles using both fuel tanks run from 650 to 750 miles, well above the range of a typical gasoline-powered pickup.

Passenger cars that run on CNG are also available. The Honda Civic GX is an NGV passenger car currently available in 37 states. Unfortunately, fewer than 4,000 were sold in 2012 due to a lack of CNG filling stations and the scarcity and expense of home fueling equipment. These can only travel about 200 miles on a tank of CNG, versus 300 miles for a typical gasoline-powered car.

These NGVs are all based on gasoline vehicle designs, with the engines and fuel systems modified to operate on CNG (Figure 1). The approach of modifying gasoline vehicles for CNG use is mainly based on manufacturing economics. Auto manufacturing plants are currently configured to produce gasoline vehicles, and because of the relatively low demand for NGVs in the U.S., it is more economical to adapt existing designs to run on both gasoline and CNG. While this approach typically adds over $10,000 to the purchase price of the vehicle, their operating costs are less than for gasoline-only vehicles. A CNG-powered vehicle gets about the same fuel economy as a conventional gasoline-powered vehicle on a gasoline gallon equivalent basis. (A gasoline gallon equivalent, or GGE, is the volume of CNG that provides the same energy as a gallon of gasoline.) Currently, CNG costs about 60% as much as gasoline on a GGE basis, meaning CNG as a vehicle fuel costs significantly less per mile than gasoline. Because natural gas burns cleaner and is less corrosive than conventional fuels, NGVs also require fewer oil changes and tune-ups, and exhaust system parts last longer. Ford, GM, and Chrysler estimate that these lower operating costs will allow heavy-duty vehicles to pay back their additional up-

Page | 72

front cost in less than two years. Similarly, Honda estimates that Civic GX owners can recoup the additional expense of the car in 2.5 to 3 years.

Figure 1. Engine and Fuel System of a Bi-Fuel NGV (courtesy AutoBlog.com)

The CNG tanks often take up additional space in the rear of the vehicle alongside the gasoline tank. The fuel delivery and ignition systems are able to switch between fuels easily.

In Europe and South America, where gasoline is expensive and NGVs benefit from a well-developed refueling infrastructure, NGVs are optimized to run on only CNG. Until a more extensive CNG refueling infrastructure in the U.S. makes dedicated, optimized NGV production economical, manufacturers are focusing on the market sector that can justify the higher initial cost of NGVs and can operate without a large infrastructure of CNG stations. Vehicles with high annual fuel usage and vehicles with a predictable daily travel pattern centered on a local fueling station – buses, heavy-duty pickups used in service fleets, sanitation trucks, and so on – are the best candidates for operating on CNG. These vehicles, along with Class 8 tractor-trailers, are currently responsible for over 70% of the annual average fuel use in the U.S. (Figure 2).

The bi-fuel heavy-duty pickups described above could potentially replace a significant portion of these vehicles. Around the U.S., service fleets for natural gas distribution companies, commuter bus fleets, sanitation trucks, and airport shuttles are already running on CNG. Several states are contracting for NGV pickups for their fleets, and companies such as Chesapeake Energy, PepsiCo, UPS, and Waste Management are purchasing NGVs or converting their vehicles to run on CNG. With this expanding market, there will be a need to ensure the quality of the CNG fuel, as discussed later in this paper. The next section describes the station equipment used to refuel CNGs and how it can potentially affect CNG quality.

Page | 73

Figure 2. Average Annual Fuel Use in the U.S. by Major Vehicle Category (U.S. Energy Information Administration, 2012)

Class-8 tractor-trailer trucks, transit buses, and sanitation trucks are responsible for over 70% of annual fuel use, and would provide the most economic benefit if converted to operate on CNG.

NGV Refueling Stations

To provide adequate driving range, CNG is stored onboard NGVs in cylinders at a pressure of 3,000 psig or (for newer vehicles) 3,600 psig. This means that natural gas provided by a local utility must be compressed at the refueling station before it is dispensed. This section describes the functions of an NGV refueling station, and also notes some of the potential causes of poor CNG quality within station equipment.

NGV stations fall into one of two types, depending upon the vehicles they serve.

• Fast-fill stations (Figure 3) are best for retail fueling of light-duty vehicles that need to refuel quickly. These stations receive fuel from a local utility distribution line at low pressure, and use an on-site compressor to increase the gas pressure to CNG levels. Once compressed, the CNG is transferred to a series of storage vessels where it is available for fast delivery to the vehicle. The storage system can be a single buffer reservoir at a service pressure around 4,300 psia, or a cascade storage system with three reservoirs at low, medium, and high pressures. A buffer storage system will refuel a vehicle in about one-third the time of a cascade storage system, but a cascade system requires less work by the compressor. The operation of a CNG dispenser is similar to that of a gasoline pump, and the dispenser is about the same size.

• Time-fill stations (Figure 4) are used primarily by fleets. In this arrangement, the low-pressure utility line delivers natural gas to a large compressor that directly fills the vehicles. A small buffer storage tank may serve to keep the compressor from cycling on and off repeatedly while fueling the vehicles. The time to fuel a vehicle can last many hours, depending on the number of vehicles and the size of the compressor. However, this arrangement provides for less heat of compression and fuller tanks than the fast-fill approach. This arrangement works well for vehicles with large tanks that refuel at a central location overnight, such as buses or sanitation vehicles. Small home refueling stations also work on this principle. Time-fill stations can be designed to fit the application (transit bus refueling, sanitation trucks, or personal vehicle), and will vary in cost accordingly, but are typically less expensive than fast-fill stations.

Page | 74

Figure 3. Example Layout of a Fast-Fill CNG Refueling Station

(courtesy U.S. DOE Alternative Fuels Data Center)

In this layout, the dryer removes water or water vapor from the natural gas before compression. Some stations place the dryer between the compressor and the storage tanks. An algorithm adjusts the CNG delivery pressure

for the ambient temperature and the heat of compression to ensure that the vehicle receives a full tank.

Figure 4. Example Layout of a Time-Fill CNG Refueling Station

(courtesy U.S. DOE Alternative Fuels Data Center)

As with the fast-fill station design, the dryer may be located before or after the compressor. A buffer storage tank is included in some stations to keep the compressor from cycling on and off during refueling.

Besides the compressor and storage tanks, both types of NGV station arrangements include a drying unit to remove excess water vapor from the fuel. Some stations are arranged to dry the gas before compression, while others compress the gas first. The dryer may use a desiccant, a molecular sieve material, or deliquescent salt tanks. If the dryer has not been regenerated or maintained properly, or if the station receives natural gas with high levels of water vapor that overload the dryer, high levels of moisture can pass through the system into the stored CNG and into the NGV fuel tank. Similarly, oiled compressors that suffer from carryover can introduce compressor oil into the CNG stream. The next section describes how these fluids can cause problems with NGV performance.

Gas Quality Concerns for NGVs

Furnaces, gas-powered turbines, and other stationary equipment fueled by natural gas can be affected by poor gas quality or changes in composition. For example, furnaces often operate with orifices sized to admit a certain amount of air and fuel for ideal combustion, producing only CO2 and H2O as combustion products (stoichiometric combustion). If the fuel gas becomes leaner or richer than the expected feed gas, the fuel-to-air ratio may no longer be ideal, leading to unstable combustion, yellow-tipping, or soot production. In the same way, an NGV engine that receives CNG that is too lean or too rich will run inefficiently and can “knock.”

NGV manufacturers typically consider two CNG quality parameters when designing engines and predicting their performance. The Wobbe number is proportional to the amount of air needed for complete stoichiometric combustion. For NGV engines and other gas-powered engines, the Wobbe number is related to engine power and the optimum fuel-to-air ratio. Deviations from the optimal Wobbe number of the fuel may result in poor

Page | 75

operational and environmental performance. The Wobbe number is equal to the higher heating value of the gas divided by the square root of its relative density (specific gravity).

Wobbe number =higher heating value

�relative density

The methane number is critical to predicting NGV engine efficiency. Fuels with higher methane numbers are better able to resist combustion knocking. The methane number for CNG is computed using the ratio of hydrogen atoms to carbon atoms in the hydrocarbon components and a parameter called the motor octane number. Different formulas can be used to compute these quantities, and the reader is referred to the American Gas Association (AGA) Gas Quality Management Manual for details.

Certain contaminants can also cause problems with NGV fuel systems or engine performance. The relative amounts of lighter hydrocarbons and diluents in CNG will affect the Wobbe number and the methane number, possibly causing efficiency problems. If large quantities of heavier hydrocarbons, such as heptanes, octanes, and nonanes, are present in the CNG, they may condense out and contaminate the fuel system, and/or may lead to incomplete combustion and poor performance. Compressor oils used in station compressors are composed of long-chain hydrocarbons (typically C12 to C15), and may pose problems similar to condensates if they enter the NGV fuel system with the CNG. Excess water vapor has been known to collect in CNG tanks, and in theory, they can form hydrates that can block NGV fueling systems. For this reason, dehydration equipment at refueling stations should be properly maintained and regularly regenerated. Sulfur compounds are also of concern because they can be corrosive to equipment, both in NGVs and in fueling stations.

Considerations for Sampling and Analyzing CNG

Both NGV manufacturers and operators of refueling stations are interested in monitoring CNG quality. Regular CNG quality data would alert station owners to changing gas compositions, compressor oils, high levels of moisture, and sulfur compounds. If problems arose, analyses of the utility gas entering the station, the CNG at the dispenser, and the gas stream at various points in the station equipment could help pinpoint the sources of any quality problems. This final section presents some thoughts on how this might be accomplished.

Natural gas entering most CNG stations is delivered at distribution pressures by the local utility company. A variety of analyzers already available to the natural gas industry can be used at this point in the stream. Some CNG stations already use in-line moisture analyzers to monitor their dehydration units, and similar devices could be used at any location upstream of the station compressors where the pressure meets the analyzer specifications. For other gas quality data, such as methane number or sulfur content, the station owner should weigh the expense of the appropriate analyzers against the benefits of regular monitoring. If the expense is not justified, a practical alternative is to collect spot samples for laboratory analysis using the methods of American Petroleum Institute (API) Manual of Petroleum Measurement Standards (MPMS), Chapter 14.1 (a.k.a., API Chapter 14.1) and Gas Processors Association (GPA) Standard 2166. If sulfur data is needed, the sampling equipment should be passivated against reactions with sulfur compounds to preserve the sulfur content of the samples.

If CNG quality is to be measured at points downstream of the compressor, such as the storage tanks or the dispenser nozzle, the pressure and temperature of the CNG can provide both advantages and challenges to accurate sampling. API Chapter 14.1 discusses the phase behavior of natural gas streams at transmission or distribution pressures, but that document provides a starting point for the discussion of CNG behavior. Figure 5, adapted from that standard, illustrates a phase diagram of a transmission gas and ways in which samples at these pressures can be distorted by phase change. Joule-Thomson (J-T) cooling through an unheated regulator or partially closed valve (path 1-2) can cause heavy hydrocarbons to condense from the sample into the equipment. J-T cooling can be offset by the application of heat to the sampling system, as shown by path 1-3. If a sample in a closed cylinder is exposed to low ambient temperatures (path 4-5), the sample may become two-phase inside the cylinder. Re-heating the sample to its original temperature for several hours before opening the cylinder will re-vaporize any liquids and restore the integrity of the sample.

Page | 76

Figure 5. A Natural Gas Phase Diagram Showing Common Sampling Processes that Can Cause Condensation and Sample Distortion (adapted from George and Kelner, 2006)

Path 1-2 represents retrograde condensation and sample distortion due to Joule-Thomson cooling through a regulator or throttle. Path 1-3 shows how adding heat through the flow restriction will avoid condensation of the

natural gas sample. Path 4-5 demonstrates how exposing a sample to ambient temperatures below the hydrocarbon dew point will cause condensation in the sample.

CNG is stored and often dispensed at pressures of 3,000 psig or 3,600 psig. At these pressures, CNG is a supercritical fluid – a single phase with properties of both liquids and gases – and well above its two-phase envelope. Moderate temperature and pressure changes would be less likely to cause condensation from the CNG than similar changes in state of transmission gas samples. Also, at CNG pressures, the Joule-Thomson coefficient is typically no more than 2°F per 100 psi, much smaller than the value of 7°F/100 psi found at transmission pressures. As a result, throttling and ambient cooling are less likely to cause phase change of a CNG sample if the sample pressure is kept well above the cricondenbar (Figure 6).

Page | 77

Figure 6. A Phase Diagram Showing the Effects of Throttling and Ambient Temperature Changes on CNG Samples

If the pressure reduction through any regulator or throttle (path 1-2) is low enough to keep the CNG sample supercritical, adding heat to avoid phase change should not be necessary. Ambient cooling of a supercritical

CNG sample (path 4-5) is also less likely to cause phase change in the sample container than similar cooling of a gas sample at transmission pressures (see Figure 5).

If the pressure of the CNG stream being sampled is stable and well above its cricondenbar, a constant-pressure sampling method should be considered for collecting spot samples downstream of the compressor. Assuming the pressure drop through any restriction is small enough to keep the sample supercritical and away from the two-phase envelope, the constant-pressure equipment would not need to be heated during sampling. However, the sample and the cylinder would still need to be kept above the cricondentherm as the sample is withdrawn for analysis, so that condensation does not occur inside the cylinder as the sample depletes.

As with any other sampling location, knowing the pressure of the supply is important for accurate sampling and analysis. CNG pressures at various locations in the system can fluctuate as CNG is dispensed or sampled, especially in fast-fill systems. Figure 7 shows pressure data taken at various locations of a fast-fill station with sequenced storage tanks. In this test, the pressure at the dispenser outlet closely followed the pressure in the NGV tank, which began the fueling cycle at 1,000 psig. At this state, the dispensed gas stream was likely below its cricondenbar. If a sample were to be drawn from the dispenser outlet at the beginning of the fill cycle, equipment heating might be necessary to keep the sample outside the two-phase envelope. Any analyzer used to collect moisture levels or other gas quality data would either need to be capable of correcting for rapid changes in pressure and temperature, or would need to be placed in a region of stable conditions.

Page | 78

Figure 7. Pressures at Various Locations in a Fast-Fill CNG Refueling Station during a Fill Cycle (from Barajas et al., 1997)

Because the pressure at the dispenser nozzle followed the pressure in the vehicle CNG tank, the gas leaving the nozzle early in the fill cycle was not supercritical. Sampling from the dispenser early in the fill cycle would require

heating or other precautions to prevent phase change and sample distortion.

A final consideration for analyzing or sampling CNG is safety. Several considerations for sampling safety are listed below; others may be discovered as the industry gains experience with this topic.

• Any in-line analyzers used to measure CNG quality must be rated for the expected pressures. Newer NGV fuel tanks are rated for a full charge of 3,600 psig, which is well above the 1,800 psig pressure rating of typical United States Department of Transportation (DOT)-rated sample cylinders and many in-line analyzers used by the transmission and distribution industry. Equipment rated for higher pressures will be necessary.

• Fast-fill CNG stations use active temperature compensation to ensure that the NGV tank is filled to this pressure after CNG temperatures stabilize, and pressures in the refueling system may briefly exceed the 3,600-psig level.

• Sampling apparatus must be fitted with relief valves or other safety apparatus to avoid over-pressurization.

• If CNG spot samples are to be transported to a laboratory for analysis, the lab must be aware of the high sample pressures, particularly if they routinely heat the samples before analysis. To comply with DOT regulations for transport, sample cylinders must also be fitted with relief valves, and must be rated for pressures 40% above the intended sample pressure – approximately 5,000 psig for CNG samples at 3,600 psig.

Conclusion

The anticipated expansion of the NGV market in the United States has led automakers and other parties to consider the quality of CNG and its potential impact on vehicle performance. In particular, NGVs may be subject to problems from changes in hydrocarbon content, moisture, and sulfur content. This represents an opportunity for the natural gas industry to help refueling station operators and NGV manufacturers monitor CNG quality, avoid possible problems, and help increase the acceptance of NGVs in the U.S. The industry’s experience with analyzing and sampling natural gas in the production, transmission, and distribution areas can serve as a starting point to develop methods for analyzing and sampling for CNG quality. The properties of CNG as a supercritical fluid may provide advantages in sampling and analysis, but the operation of CNG refueling stations and the safety requirements of working with pressures above 3,000 psig must also be considered.

Page | 79

References

American Gas Association, AGA Gas Quality Management Manual, Washington D.C., USA, August 2013.

American Petroleum Institute, API Manual of Petroleum Measurement Standards, Chapter 14 – Natural Gas Fluids Measurement, Section 1 – Collecting and Handling of Natural Gas Samples for Custody Transfer, Sixth Edition, Washington D.C., USA, February 2006.

Autoblog Green Car News, http://green.autoblog.com/category/natural-gas.

Barajas, A. M., Svedeman, S. J., and Buckingham, J. C., “NGV Fueling Station Technology Program,” Gas Research Institute Report GRI-97/0112, published by Gas Technology Institute, Des Plaines, Illinois, USA, July 1997.

Bowles, E., “CNG Going Forward (as a Transportation Fuel),” presentation to the North Texas Measurement Association, October 18, 2013.

George, D. L., and Kelner, E., “Additions and Changes to the Latest Revision of API Chapter 14.1,” Proceedings of the Eighty-First International School of Hydrocarbon Measurement, Oklahoma City, Oklahoma, USA, May 2006.

Gas Processors Association, GPA Standard 2166-05, Obtaining Natural Gas Samples for Analysis by Gas Chromatography, Tulsa, Oklahoma, USA, October 2005.

James, G., “The Next Big Thing,” American Gas Magazine, March 2013.

Natural Gas Council and NGC+ Interchangeability Work Group, White Paper on Natural Gas Interchangeability and Non-Combustion End Use, February 2005.

Kubesh, J., King, S. R., and Liss, W. E, “Effect of Gas Composition on Octane Number of Natural Gas Fuels,” SAE Technical Paper 922359, October 1, 1992, http://www.sae.org.

Stenquist, P., “The Case for NGVs,” American Gas Magazine, October 2013.

U.S. Code of Federal Regulations, Title 49, Section 173.301, General requirements for shipment of compressed gases and other hazardous materials in cylinders, UN pressure receptacles and spherical pressure vessels, U.S. Department of Transportation, October 1, 2012.

U.S. Department of Energy, Alternative Fuels Data Center, http://www.afdc.energy.gov.

U.S. Energy Information Administration, Annual Energy Outlook 2012, with Projections to 2035, Report DOE/EIA-0383(2012), June 2012.

Wiley, A., and Hunt, T., “CNG as Vehicle Fuel Looming Larger,” Pipeline & Gas Journal, November 2011.

Page | 80

NEW TECHNIQUES in LNG SAMPLING

Ken Thompson, Mustang Sampling

Introduction

Transportation of LNG (Liquefied Natural Gas) has increased with many new export and import terminals that have been built to date and more under construction or in the planning stages due to the increase in world demand of LNG. This demand has increased due to LNG being considered by many countries as the fuel of choice based on its safe properties, availability, cost, cleanness and ease of transportation. With the increase of liquefaction, storage, blending and transportation so has the need increased for “Analytically Accurate” measurement of the LNG in the realm of custody transfer. This has also led to new techniques in LNG sampling to help verify the measurement results. This paper will cover these new techniques in their approach and their unique challenges for the industry.

Standards for Sampling of LNG

There are several standards/publications that cover the sampling of LNG.

The primary standard used is the International Standard – ISO 8943 (Second Edition)

• ISO 8943-2007 – Refrigerated light hydrocarbon fluids – Sampling of liquefied natural gas – Continuous and intermittent methods.

• First edition was published in 1991, second edition in March 2007.

• The standard specifies methods for the continuous and the intermittent sampling of LNG, specifically during transferring through a LNG piping system. The standard takes an in depth look at the sampling system components to include the sample probe, sample vaporizer, sample holders and sample cylinders.

The second reference standard is the GIIGNL – LNG Custody Transfer Handbook (Third Edition)

• GIIGNL (Groupe International des Importateurs de Gaz Naturel Liquefie)

• First edition was written in 1987, second edition in 2001 and third edition in March 2010.

• The methods described in the handbook are given to help improve existing procedures and can also be used in purchase and sales agreements or serve as a reference for new agreements. The handbook is based on the measurement methods most used by GIIGNL members.

There is a third publication (pre GIIGNL & ISO 8943) that is the National Bureau of Standards (N.B.S.) “LNG Measurement – A User’s Manual for Custody Transfer” – 1985 edition.

Additional standards include:

• ISO 10715-2001, Natural Gas sampling guidelines.

• API 14.1-2006, Collecting and handling of natural gas samples for Custody Transfer.

• BS EN ISO 12838-2001, Installations and equipment for liquefied natural gas-suitability testing of LNG sampling systems.

The LNG Sample System Overview

The objective of the LNG sampling system is to provide consistent, repeatable analyzed results during vessel loading or unloading of LNG cargo into or from a pipeline during stable flow. The “Stable Sampling Time” is the most important point during the LNG measurement cycle and all of the sampling procedures must be consistent during the entire operation to provide consistent and repeatable results. Samples included in a total report that are outside of the “stable sampling time” provide the majority of the errors that occur in the final calculation of data. Any sample analysis taken during interruptions to temperature, pressure or LNG flow during the “stable sampling time” should be removed from the final analysis results.

Page | 81

The sampling and analysis of LNG must be performed at both the loading and unloading terminal. These subsequent analysis data will reveal differences between loading and unloading reports due to the “Weathering” of the LNG or as some refer to it as the “ageing”. This phenomenon occurs over time due to the effects of “boil-off” which happens when the lighter components of LNG vaporize and exit the cargo tank.

• “Aging and/or Weathering” – terms used synonymously to describe the ongoing process of “boil-off”.

• LNG boil-off – as the cargo gains heat, the lighter components of the LNG exit the cargo tank and is often used as ship’s fuel.

Since the weathering effect alters the cargo chemistry it also alters mass and the monetary value of the cargo. The design and accuracy of the LNG sampling system will have a direct effect on the calculated density and gross calorific value of the LNG being transferred and the final financial value of the cargo. A typical LNG ship value range is 50 – 75 million US dollars, therefore a small error of 1% in the energy calculation can equal $500 to $750 thousand dollars error during custody transfer of even a single cargo. Sampling Process Overview The sampling process that has been used in the US import terminals has been a combination of two types;

• Continuous – consist of a sample vaporizer system and a gas chromatograph on a continuous basis during a constant stable flow rate.

• Spot (also referred to discontinuous sampling in some publications) – consist of sample cylinders (normally 3 ea.) that are taken manually or automatic at predetermined intervals during custody transfer of the LNG (this should also be during the “stable sampling time” discussed earlier).

The sampling process that is normally used outside of the US by export and import terminals consist of a much broader procedure that includes multi sampling techniques and multi gas chromatographs analyzing the LNG cargo at various stages throughout the custody transfer. The trend that I see going forward with the advent of our export terminals will be to closely follow the European practices and the standards, as referenced earlier which will constitute change in domestic procedures. The sampling process includes the following; Ship (vessel) Loading:

• Continuous – consist of a sample vaporizer system and a gas chromatograph on a continuous basis during a constant stable flow rate as described above.

• Spot (also referred to discontinuous sampling in some publications) – consist of sample cylinders (normally 3 ea.) that are taken manually or automatic at predetermined intervals during Custody Transfer of the LNG (this should also be during the “stable sampling time” discussed earlier).

• Composite – an automatic collection system that over the duration of the “stable sampling time” can provide an “averaged” sample of the pipeline contents, therefore eliminating the potential risk of biased manual spot sample. The system to include a sample pump or pumps (computer controlled) with a 1-5 liter gas sample holder (water-seal type or waterless type) and 2-6 sample cylinders (manually filled) for transferring to a lab GC for analysis.

• Cargo sample cylinder – this is a cylinder of gas (300-500 CC’s) that is taken during the loading of the ship and is transported with the ship to the unloading terminal for analysis by the unloading terminal’s lab GC.

Ship (vessel) Unloading

• Continuous – consist of a sample vaporizer system and a gas chromatograph on a continuous basis during a constant stable flow rate as described above.

• Spot (also referred to discontinuous sampling in some publications) – consist of sample cylinders (normally 3 ea.) that are taken manually or automatic at predetermined intervals during custody transfer of the LNG (this should also be during the “stable sampling time” discussed earlier).

• Composite – an automatic collection system that over the duration of the “stable sampling time” can provide an “averaged” sample of the pipeline contents, therefore eliminating the potential risk of biased manual spot sample. The system to include a sample pump or pumps (computer controlled) with a 3-5 liter gas sample holder (water-seal type or waterless type) and 3-5 sample cylinders (manually filled) for transferring to a lab GC for analysis.

Page | 82

• Cargo sample cylinder is analyzed by the unloading terminal lab gas chromatograph.

• The Certificate of Quality that is supplied for each ships LNG cargo is compared to all four of the above sampling procedures results and if within a predetermined range “accepted or rejected”.

It is worth noting that at almost all terminals there are redundant LNG sampling systems installed for the purposes of having a hot online standby system and in most cases are ran simultaneously during loading and unloading and the results of each system also compared to the Certificate of Quality provided with the cargo. The LNG Sampling System A LNG sample system consist of the followings items:

• Sample probe with isolation valve – The sample probe must be installed at a point in the LNG pipeline where the degree of sub-cooling is high. The preferred installation is on a Horizontal plane in the pipeline. Sub-cooling must be maintained to prevent pre-vaporization (formation of bubbles) of the LNG prior to entering the vaporizer. This is achieved by properly insulating of the sample probe and isolation valve.

• LNG sample probes are constructed of stainless steel.

• Vacuum jacketed sample tubing (VJT) – In some cases SS tubing is used as the liquid sample line and is insulated, however over time the insulation will become wet and saturated and will lose its insulating value. The best liquid sample line is VJT with a small inside diameter (small diameter tubing acts as a restriction orifice prior to vaporizer) and as short of a length as possible (recommended not to exceed 7 meters).

• Inlet LNG temperature sensor – To ensure the liquid LNG is in the proper sub-cooling region you should monitor the temperature at the inlet to the vaporizer, data log this information and have alarm set points in case the temperature is not in proper region.

• Sample vaporizer coil or coils and control system – The vaporizer must be designed to avoid fractionation of the liquid LNG and be capable of heating to a sufficient high temperature, 50 deg. C or greater to ensure immediate vaporization of even the heaviest trace components. The vaporizer shall be constructed so the heavier components of the LNG will not remain in the vaporizer (flow should be top entry and bottom exit).

• Vaporizer outlet temperature sensor – Ensures proper exit temperature from the vaporizer and provides data logging and alarm set points.

• Inlet pressure sensor – Monitors the inlet pressure of the LNG as it exits the vaporizer. Monitoring, data logging and alarm set points ensures the pressure is above the liquid-vapor equilibrium curve.

• Accumulator/Mixing chamber – Ensures a well-mixed (homogeneous sample) after vaporization of the LNG. Accumulator/mixing chamber includes a mixing wand (inside the accumulator) and a sample probe (inside the accumulator) for the gas exiting to prevent taking sample from the wall of the accumulator/mixing chamber where heavies or contaminants may be trapped. Chamber to be insulated.

• Vaporizer flow control/regulator system and control system – Maintains even flow through the vaporizer coils regardless of the inlet pressure and ensures a constant flow volume to provide a fresh sample for analyzing when required for the on-line gas chromatograph, composite sampler or spot sample cylinders.

• Outlet heated pressure regulator – Pre & Post heated regulator maintains a constant outlet temperature of the natural gas preventing any condensates in particular hydrocarbon vapor condensation. Outlet temperature should be 38 deg. C.

• Outlet gas temperature sensor – Ensures proper outlet temperature and provides data logging and alarm set points.

• Outlet pressure sensor – Ensures proper outlet pressure to gas chromatograph and provides data logging and alarm set points.

• Sample holders and sample cylinders – For the composite sampler and spot sample requirements.

• Heat trace sample tubing – For transporting heated gas sample from the outlet heated regulator to the gas chromatograph.

• Gas Chromatograph – Analyzes the vaporized LNG as a vapor gas.

• Data Acquisition and Processing System – Provides data logging/alarming for the LNG sample conditioning system and gas chromatograph. Provides Loading and Unloading Reports.

Page | 83

Conclusion The analysis of LNG is a complicated procedure due to the transported temperature and the loading and unloading temperature and pressure being so close to the boiling point. (-160 deg. C & 1.3 bar). The preferred loading and unloading pressure is above 3 bar which results in an increase of the boiling point to -146 deg. C. for most mixtures. There are many factors that can contribute to heat gain in the LNG terminal piping system and the LNG sampling system that will directly contribute to errors in the analysis. Therefore the system must meet or exceed the standards listed in this paper. Provided the user has installed the sample system properly, maintained it correctly and obtained consistent and repeatable results, then they should trust the system to be “Analytically Accurate”. References

(1) ISO 8943-2007

(2) GIIGNL-2011

(3) N.B.S. “LNG Measurement – A User’s Manual for Custody Transfer”

(4) API 14.1-2006

(5) Kenbar. A. (2012) Assessment of LNG sampling systems

Ken Thompson 12/2013

Page | 84

IMPROVING THE SPEED AND ACCURACY OF WATER VAPOR AND HYDROGEN SULFIDE

MEASUREMENTS BY OPTIMIZING THE SAMPLE TRANSPORT SYSTEM.

Phil Harris, HariTec

John Kozich, O’Brien Corporation

Keywords: Sample Systems, Sample Transport, Adsorption, Desorption

ABSTRACT

Natural gas is processed in order to meet customer specifications, such that it may be sold based on energy

content. Suppliers and producers whose product does meet the customer / pipeline specifications are “shut-in”

until it can be demonstrated that the product is within specification. With the advent of shale gas production and

more complex and diversified streams entering the transportation, there is increased interest and demand for

accurate, reliable and timely analysis of gas quality. The validity of such measurements can be strongly impacting

by the sample gas transport system, the process by which the gas is removed from the process and move to the

analytical systems. For trace components like moisture and hydrogen sulfide, the adsorption and desorption of

the analyte from surfaces in the system must be considered. A thorough description of the effects as well as

recommendations on system optimization is presented.

INTRODUCTION

The analysis of natural gas has various purposes, from process optimization, to ensuring critical specifications are

met during custody transfer, to determination of product value. In each of these cases the accuracy and

timeliness of results are of paramount importance. While the response speed of the analytical technique has

important implications here, it is well accepted that the design and implementation of the analyzer sample system

that often determines the overall performance of the system.

Effective sample conditioning and transport is imperative to providing a representative sample of process fluids to

an extractive process analyzer1.The process of sample conditioning begins at the sample point, where preferably

a probe is used to extract a sample of the gas. Frequent sample conditioning operations performed at the probe

include filtration for particulate and/ or liquids and may include pressure control. External to the probe there may

be additional pressure reduction operations which often must be performed under controlled temperature

conditions, flow control and additional filtration. The pre-conditioned sample is then transported to the process

analyzer, where additional steps may be taken to remove contaminants which may alter the sample composition

or cause problems with the analysis. Care must be taken to perform each of these operations in an optimized

manner if we are to expect a meaningful result from the process analyzers.

Herein, we intend identify key concepts which must be addressed when designing a proper sample system, with

an emphasis on one of the least known and considered effects – the adsorption and desorption of analytes in the

sample transport tubing and its effect on accuracy and response time.

SAMPLE CONDITIONING

Sample conditioning is the process of extracting a “representative” sample from a process pipe or vessel, making

it compatible with the sample transport system, and further treating it such a way as to make it suitable for the

analyzer or analytical technique chose. The sample conditioning system is a physical assembly of fluid

processing components that ensures that the sample delivered to the analyzer is compatible with its requirements

despite any and all process fluctuations. It exists as a physical entity within the much larger environment of the

Page | 85

sample handling system, which encompasses to some extent the process itself as well as the sample extraction,

transport, conditioning, disposal arrangements and the analyzer itself2.

The purpose of the sample conditioning system is to deliver a representative sample to the analyzer for

subsequent analysis. In terms of natural gas processing, numerous definitions of representative sample exist3.

From the Gas Processors Association publication GPA 2166-05, "The objective of the listed sampling

procedures is to obtain a representative sample of the gas phase portion of the flowing stream under

investigation. Any subsequent analysis of the sample regardless of the test is inaccurate unless a

representative sample is obtained.” And from ISO-10715 a representative sample is, “A sample having

the same composition as the material sampled, when the latter is considered as a homogeneous whole.”

Finally, API 14.1 offers a similar statement in the latest revision, “a representative sample is

compositionally identical or as near to identical as possible, to the sample source stream.” These

standards are the most common and current ones referenced on gas sampling procedures.

The single largest source of error in the analysis of natural gas samples is distortion of the sample composition

while extracting, transporting or conditioning the sample for the analyzer4. In general, it is assumed that sampling

clean dry natural gas which is well above its hydrocarbon dewpoint is simple, while sampling natural gas that is

near its dewpoint is much more problematic. This is arguably true for bulk compositional measurements such as

hydrocarbon composition and energy content, but may not be true when one considers trace components such as

sulfur species and water content. In the latter cases, material compatibilities must be considered carefully and the

choices made may significantly affect the analytical results, even for clean dry gas samples.

In evaluating and selecting materials, many guidelines and regulations exist to assist us in matters such as

corrosion resistances (NACE), or in regards to elastomer compatibility and seals. Less often addressed is the

subject of adsorption and desorption of trace components in the sample handling system. Adsorption desorption

effects are most relevant during the transport and analysis of trace components: at high concentrations the

surfaces quickly come to equilibrium and the surface is not able to substantially change the concentration of the

gas but at low concentrations the surface may can absorb a significant percentage of the component present and

may take a long time to come to equilibrium.

In all cases, the delivery of a representative sample begins with extraction of the sample gas from the process,

and is followed by the subsequent transport of said gas to the analyzer.

SAMPLE GAS EXTRACTION

All gas samples should be extracted through a sample probe and this component must be considered as the first

part of the sample transport system. A commonly applied rule of thumb is that the probe should extend into the

central 1/3 of the process pipe, although this decision may be impractical in some cases due to probe resonance

effects 5, 6. Sample probes within a flowing pipeline can vibrate as the gas forms eddy behind the probe. If the

probe is too long, the vibration frequency can eventually cause the probe to break.

In extracting the process sample, one must consider whether in-situ filtration and / or in-situ pressure reduction

should be applied and the impact of these operations on producing a representative sample. This is well

addressed in API 14.1 and has been thoroughly covered by other authors 7, 8. Filtration to prevent the further

transport of particulate and especially liquid droplets should be performed at the probe and at pipeline pressure

and temperature. This fact clearly indicates the benefits of in-situ filtration probes in all cases. Changing the

temperature or the pressure of the gas before eliminating entrained liquids will always change the gas

composition.

In sharp contrast, in-situ pressure regulation should only be employed in cases where it is known sufficient

dewpoint margins exist. As a safety margin against uncertainties in predicted hydrocarbon dew points, API

Chapter 14.1 recommends that sampling equipment be maintained at least 30ºF (17ºC) above the predicted

hydrocarbon dew point. In situations where the gas is at or near its dewpoint in the pipeline, an in-situ regulator is

Page | 86

generally unacceptable as the regulator will be at the same temperature as the pipeline and provide insufficient

excess heat energy to prevent condensation occurring during the expansion.

SAMPLE TRANSPORT AND SAMPLE LINES

Specifying the best sample line for a project requires a thorough analysis of the system and requirements. Of

particular importance are:

1) Species to be analyzed,

2) Sample gas composition and dewpoint / phase behaviour,

3) Length of the sample line run,

4) Operating pressure and temperatures,

5) Required gas velocities and response times, and

6) Material compatibility.

The bulk composition and phase behavior are required to determine the operating temperature necessary to

prevent sample condensation, dewpoint issues. Recall that API 14.1 recommends all equipment be maintained at

least 30ºF (17ºC) above the worst case dewpoint temperature. Assuming that heating is required for the line –

there is little benefit in providing only marginal capabilities and thus a larger dewpoint margin is warranted.

The length of the sample run, pressures, temperatures and required response time all play important roles in

specifying the tubing diameter and gas flow rates. Care should be taken to also consider the pressure drop during

the transport of the gas, although gas phase pressure drops are typically small at the flow rates used for analyzer

sample systems.

Materials compatibility can become a critical factor in terms of analytical performance, accuracy and response

time. The sample line often represents the largest available surface area for gases to adsorb and desorb in the

flow path, and thus is a critical element in ensuring a representative sample is delivered to the analyzers. The

adsorption desorption process can result in long response times or even completely erroneous values when

measuring trace species such as hydrogen sulfide, water vapor or mercaptans.

The surface chemistry of the sample line is of paramount importance in determining the adsorption / desorption

characteristics and rates. Pressure affects molecule density and linear flow rates through the line, and thus affects

the kinetics in a number of ways. Temperature can greatly affect desorption rates. All of these parameters must

be understood to provide physio-chemical model of the mass transport in sample lines.

ADSORPTION DESORPTION EFFECTS

When a sample of gas touches the walls of any tubing or container, some of the molecules stick to the surface.

The surface contains a multitude of adsorption sites where the molecules may stick. Imagine a metal surface that

has been scrupulously cleaned, and all these active adsorption sites are available. These active sites are surface

structures where an uneven distribution of electrons causes negative or positive charges to accumulate 9. When

we allow a gas containing trace amounts of polar molecules such as H2S or H2O to contact this surface, the

molecules flock to these activate sites. As the number of molecules stuck to the surface increase, fewer sites are

available and the rate of adsorption decreases. As well, molecules stuck to the surface occasionally escape

(desorb) and come back out into the gas phase. The system reaches equilibrium when the rate of adsorption and

the rate of desorption equal each other.

There exists a popular (if ill-conceived) notion that once this effect has occurred the first time, the line has been

conditioned or pickled and no more adsorption or desorption will occur. However, this is not the case. If the

concentration of the analyte increases, the rate of adsorption increases and the system must now achieve a new

equilibrium. If the concentration of the analyte suddenly drops, then the rate of adsorption decreases but

desorption stays the same – and the concentration at the analyzer slowly tails to zero.

Page | 87

Material treatment and coatings can greatly affect the rate of adsorption loss to surfaces. This effect has been

extensively demonstrated for sulfur compounds in both static (sample cylinders) and dynamic (sample line)

applications. In sample cylinders, complete loss of sulfur species can be seen in a few days for uncoated

vessels, but coated vessels can retain sample integrity for days or weeks for some species as depicted in Figure

1. 10 In flowing systems, the adsorption effects substantially delay the appearance of the inlet gas at the exit of the

tube as shown in Figure 2 11, where the 316L stainless steel line does not come to equilibrium in 15 minutes when

challenged with 0.863 ppm H2S.

Figure 1 Samples collected in Sulfinert® treated cylinders are significantly more stable than those

collected in untreated cylinders. Samples in untreated cylinders quickly lose reactive sulfurs, due to

interaction with the stainless steel surface.

Given the obvious and dramatic impact of surface treatments on the adsorption of gases of interest in the natural

gas, refining and petrochemical industry, it is important that we understand first what how these treatments affect

the surfaces involved and how those effects impact the mass transport. The mass transport properties can be

examined in varying degrees of mathematical rigor, and we have chosen to develop a model which encompasses

the major effects in evaluating transport along sample tubing.

The model proves useful to evaluate the effects of changing parameters like pressure, temperature, flow rate and

tubing diameter on response time in analytical systems.

Page | 88

Figure 2 Flowing H2S in nitrogen sample – 6 foot sample line at a linear velocity of 2 cm/sec.

SURFACE MORPHOLOGY

The chemical and structural characteristics of a sample transport tube determine to a large extent how the tube

and the sample it is transporting will react. The chemical composition of the metal obviously affects corrosion

rates, and it should be equally apparent that changes in the chemical composition will affect the type of chemical

reactions and the binding energy of adsorbed molecules on the tube surface of a given metallurgy. The surface

roughness of the tube walls greatly affects the surface area available for the reaction to occur on – smooth mirror

like surfaces offering substantially less surface as compared to rough walls. Various surface treatments affecting

the surface chemistry or roughness or both are used to prepare stainless steel tubes for use in sample transport.

The interior surface of a stainless steel tube can be of variable quality and composition and is dependent on the

initial material quality, fabrication methods and post fabrication–processing. The most widely specified material for

instrument sample lines and gas distribution applications is 316L stainless steel12. This 316L stainless is able to

form a stable and protective oxide layer primarily due to the presence of more than 16% chromium and the

formation of stable and inert chromium oxide (Cr2O3) on the surface13. Despite this fact, there is significant iron

content in such tubing, and typical chromium to iron ratios are less than 1.5:1. Indeed, many “as-manufactured”

tubes will have free iron on the surfaces, which is highly reactive. The presence of free iron is often determined

through the use of ASTM A380 “Ferroxyl Test for Free Iron”. The iron surface and iron oxides are readily

attacked, and provide pathways for both corrosion and increased surface activity under harsh conditions14.

Furthermore, the iron oxides provide active adsorption sites which strongly absorb species such as water15 and

hydrogen sulfide16.

To address such issues, as-manufactured tubing is often subjected to post-treatment in an effort to improve the

performance of tubing in real-world applications. Such post-treatment may include:

Page | 89

A) Chemical passivation, B) Electropolishing, and/or C) Coating.

Chemical cleaning and passivation treatments on stainless steel tubing are an important aspect in preparation of

these surfaces for use in critical environments and applications17. The surface to be passivated must be clean

and oil free18. Passivation is the removal of exogenous iron or iron compounds from the surface of stainless steel

by means of a chemical dissolution, most typically by a treatment with a dilute nitric or citric acid solution that will

remove the surface contamination, remove iron, enhance the chromium to iron ratio, but will not significantly affect

the stainless steel itself. Furthermore, the chemical passivation oxidizes the surface and the ratio of chromium

oxide to iron oxide increases even more dramatically. In addition, the acid will tend to attack high points and sharp

peaks more quickly than smooth portions of the tube. As a result, chemical passivation of the tube also reduces

the surface roughness. Note that surface roughness is most commonly expressed in terms of Ra, which is a

measure of the mean deviation from the centerline or average height expressed in micro-inches.

Electropolishing is often performed on as-manufactured tubing of the highest quality. Such tubing meets strict

compositional guidelines and starts with a nominal surface roughness of 20 Ra .The process uses a mixed acid

solution as the electrolyte - and a cathode is drawn through the inside of the tube. The tube becomes the anode,

so it preferentially dissolves free iron, removing metal from the peaks.19

Electropolishing further increases the chromium to iron (Cr/Fe) ratio on the surface and oxidizes the chromium to

greatly enhance the chromium oxide to iron oxide (CrOx/FeOx) ratio, and reducing the reactivity of the surface.

The process reduces the surface roughness, resulting in a mirror polished surface with much less surface area

exposed for molecular reactions to occur. In addition to appearance, electro polished tubing has five primary

advantages24:

1) Extremely smooth surface, which minimizes adherence of particles and adsorption of gases or liquids,

2) Removal of all oils and iron from the surface, 3) Increased chromium to iron ratio which improves corrosion resistance and reduces chemical activity, 4) Creation of a passive chromium oxide layer that is free of iron contamination, 5) Improved mechanical property performance through minimization of surface stresses.

In addition, to electropolishing, it is possible to preserve, protect and enhance the performance of stainless steel

tubing by adding an inert coating. In particular, it has been demonstrated a mechanically robust and long-lasting

coating can be produced through the deposition of an amorphous silicon layer onto, and into, the steel surface via

a chemical deposition process20 at 400°C. The process can be further enhanced to improve surface inertness and

reduce moisture hold-up21. The initial surface roughness of the electropolished tubing is approximately 7-10

micro-inches, to which a 5 micron coating of amorphous silicon is deposited and further chemically treated to

increase inertness and hydrophobicity.

Typical microrgraphs of stainless steel tubing that has been subjected to various surface treatments are shown in

Table I. In general, the surface roughness decreases dramatically as one looks from left to right in the table. The

passivation and electropolishing processes eliminate free iron and greatly increase the chromium to iron ratio at

the surface. Note that the electropolishing process can reduce the surface area available for chemical reaction by

as much as an order of magnitude24 .The coating processes (SilcoTech®) provide a stable, inert passivation layer

with no exposed metal oxides. This important modification to the surface chemistry does not necessarily reduce

the rate at which molecules adsorb, but greatly reduces the energy with which they bind to the surface and thus

the molecules desorb easily and remain primarily in the gas phase.

TABLE I. TYPICAL SURFACE MICROSCOPIC IMAGES OF STEEL TUBING

Page | 90

Conventional 316

SS

SilcoSteel on

Conventional

Chemically

Passivated

Electropolish Electropolished

And Sulfinert

CHROMATOGRAPHIC ANALOGY

It is useful to compare the transport of analyte molecules of interest in an extractive analyzer installation, (or the

transport of gases in a gas distribution network) to the events that occur in a gas chromatograph which employs a

capillary column. In a gas chromatographic, a small sample of analyte is injected and transported along capillary

column by the inert carrier gas. Capillary columns may have typical dimensions of 0.1 mm internal diameter and

10 meter length. Thus, the line length to ID ratio is about 10,000 to 1. In an analyzer installation, we transport

analyte molecules of interest (e.g. H2S) in an “inert” carrier gas (e.g. methane) along a sample line with typical

dimensions of 0.18 inch id and 150ft lengths, giving a length to ID ratio of 10,000 to 1! As gas is transported down

the column (sample line) – it is adsorbed and desorbed from the stationary phase (wall) and the appearance of

the gas at the end of the column (sample line) at some time which is significantly delayed from the time the carrier

gas first exits.

This analogy is useful – in that it allows those familiar with chromatography to directly apply their intuitive

understanding of that field to sample lines and mass transport. Common rules of thumb apply – for example – a

rough dirty surface in a column will lead to peak tailing or no peak at all (the effect of using standard drawn

stainless tubing in some applications). Increasing the temperature of the transport line will increase the response

speed – and temperature needs to be tightly controlled. Clean smooth surfaces make for more inert, better

responding columns (i.e. electropolishing). For the best response, a chemical inert coating should be applied to

the column (amorphous silicon coatings).

ADSORPTION, DESORPTION AND MASS TRANSPORT

The surface of stainless steel tubing is a mixture of oxides of the various compounds that make up the steel. For

simplicity, we will consider it a mixture of two types of sites, surface sites that won’t absorb an analyte of interest,

and surface sites that will. In Figure 2, we depict sites that are able to absorb a molecule of water as brown iron

oxide lattice structures, and sites that will not adsorb a water molecule as green chromium oxide lattice structures.

This is an artistic representation and in fact water can potentially adsorb on both surfaces. In fact, it has been

shown that water adsorbs on at least five different types of sites in stainless steel22. Equally important, it is the

interstitial spaces at grain boundaries which often act as traps for adsorbed species. Although water vapor is

depicted as the adsorbed species in Figure 3, the Figure applies equally well to other chemical species.

The rate of adsorption out of the gas phase and onto the surface is proportional to the concentration of adsorbate

molecules in the gas phase and the number of free sites on the surface23, thereby following Langmuir isotherms

and kinetics. It should be immediately apparent that processes such as electropolishing which reduce the amount

of surface area (and thus the number of adsorption sites) will reduce the rate at which the adsorbate molecules

get adsorbed and the total amount of molecules the tube can adsorb. It is important to also realize that in most

cases, the adsorbate can spontaneously be released from the surface as well, with the rate of desorption being

Page | 91

proportional to the number of adsorbed molecules. It is the competition between these two processes that

determines the maximum amount of adsorbate the tube can hold at equilibrium.

Referring to Figure 4, we can use the representation shown to begin to consider a model for the adsorption /

desorption processes, and to further develop that model to provide useful predictions of the mass transport

phenomena that occur as a reactive gas flows down a tube. Such systems are typically solved through partial

differential equations, which for this system may be represented as:

(1)

Figure 3. Water adsorption on stainless steel surfaces

However, such representations provide little understanding to the layman and the solution of such equations

requires sophisticated numerical analysis packages. Rather than pursue this approach, we propose to solve the

mass transport problem using a series of simplified finite difference equations, similar to the approach taken by

Air Products25, and which can then be easily implemented in a spreadsheet such as Microsoft ExcelTM.

The sample tube of length L is divided into a large number of individual elements, each of length �l, internal

radius r, surface area, SA (SA=2�r�l) and volume V (V= 4�r2�l/3). The gas is flowing into the tube at flow rate

F, so moves down the tube with velocity v, where v= F/��r2. Thus, the gas will pass through the volume element

�l in a time �t=��l/v.

The gas concentration flowing into the first volume is the inlet concentration or the concentration with which we

are determining the mass transport characteristics, and this gas is allowed to flow into the tube at pressure, P. We

will assume that at the flow rates we are working with that P does not change substantially along the length of the

tube, and as well that the tube is maintained at some constant temperature, T.

Page | 92

It will be assumed that the tube has some number SI active sites per square centimeter of linear surface area for

adsorption available initially (before any have been occupied by adsorbed molecules), so the total number of sites

in a volume element is SI multiplied by the surface area of the element, SA.

Figure 4. Mechanism and Elements of the Model

The concentration of analyte molecules that exits volume element i of the tube between time t and time t+��t will

be given by Equation 2.

Ci+1(t+�t) = Ci(t) - kads•Ci(t)•Si(t)•SA + kdes•[ SI-Si(t) ]•SA (2)

In Equation 2, Ci(t) represents the concentration of analyte molecules flowing into segment i so the first term

represents the inlet concentration. During this time interval, gas phase molecules,Ci(t), react with available

surface sites,Si(t), and get adsorbed with some rate constant kads. Thus the second term in (2) represents the rate

at which molecules are adsorbed out of the gas phase and onto the surface. In the third term, the difference SI –

Si(t) represents the number of sites which are currently filled with analyte molecules (recall that SI is the total

number of sites that could possibly hold a water molecule with Si(t) represents the number of free sites available

in volume element i at time t).

While Equation 2 gives us a finite difference equation in the gas phase concentrations to work with, a similar

equation for the adsorbed phase is needed as well, and is presented in Equation 3.

Si(t+�t) = ( Si(t) - kads•Ci(t)•Si(t) + kdes•[ SI-Si(t) ] ) •SA (3)

The rate constant for adsorption, kads, is typically relatively independent of the tube material and is primarily

dependant on collision frequency with the walls. However, how long it stays on the surface is strongly dependent

on the surface chemistry. Thus, the rate of adsorption will usually increase with increasing temperature since the

kinetic theory of gases predicts collision frequency being proportional to T1/2. However, the dominant effect with

increasing temperature is a rapid increase in the desorption rate, kdes. The desorption rate constant typically

Page | 93

follows an Arrhenius behavior, and thus increases exponentially with increasing temperature. This desorption

rate is expressed as:

kdes(T) = A•e-�E/RT (4)

�E is the activation energy to break the bond of the adsorbed state, R is the Ideal Gas Constant, T is the

temperature and A is constant. This provides us some insight into the effects of chemical passivation and

chemically treating the tube wall. Such treatments change the surface chemistry and produce a substrate where

the adsorbate (such as water) bonds weakly and thus desorbs more easily at a given temperature.

As mentioned previously, rough surfaces have large surface areas per unit length of tube, and thus have larger

numbers of free sites, SI. This increases both the rate of adsorption and the total amount of adsorbate the tube

can hold. Surface treatments such as chemical passivation and electropolishing reduce the surface area and

thereby the number of free sites. Furthermore, such treatments change the surface chemistry by converting the

strongly adsorbing iron oxide rich surface to a weaker adsorbing chromium oxide rich surface. Again, the

adsorption rate may be similar, but the retention time or the mean time spent on the surface can be quite different.

Weaker adsorption sites have a lower activation energy (see Equation 4) required to break the adsorbate-surface

bond, and thereby increases the desorption rate. Similarly, chemical treatments such as the application of an

inert glass-like layer on the surface of the tube further reduce the bond strength and increases desorption.

RESULTS AND DISCUSSION

Recently, there have been numerous studies on the mass transport of trace species through sample lines of

various compositions. Such studies have been performed with analytes such as water vapor26, hydrogen sulfide27,

and methyl mercaptan28.

The water vapor data was obtained for nominal concentrations of 1 ppm and using 100 foot sample lines operated

at 60 °C. The data clearly shows the effect of changing the material used in the sample line, with electropolished

(EP) and electropolished/Silconert lines (EPS) demonstrating much faster wet-up and dry-down times in the test.

It is clear that chemical treatment such as electropolishing or applying amorphous coatings dramatically affects

the mass transport characteristics and the suitability of the tube for transporting samples of industrial interest.

Example “wet-up” data is presented in Figure 5. In this figure, the concentration of water at the exit of the tube

was monitored as a function of time after a step change in water concentration was injected into the tube.

The theoretical model described previously was implemented in Excel®, and used to simulate the results obtained

during empirical testing. The Model results are shown in Figure 6. While the model in its present state does not

predict the results obtained empirically with great precision, it definitely identifies the common trends of the data.

Further refinement of model parameters is required. Of greater importance, the use of the model now allows us

to extrapolate the empirical results obtained. In the model, we can readily change the pressure, flow rate or

sample line length and observe the effects on response speed.

Page | 94

Figure 5. “Wet-up” Data for 1 ppm challenge gas at 0.35 slpm

Figure 6. “Wet-up” Model for 1 ppm challenge gas at 0.35 slpm

Assuming that the model is deemed valid, the use of the model allows us to also extrapolate the data to different

inlet concentrations. Such extrapolations are of course extremely useful, as they alleviate the requirement to

repeat experiments at a variety of different conditions, and allow for rapid evaluation of alternative solutions.

Page | 95

Figure 7. Response TIME to a 100 ppb “wetup”

Figure 8. Response to a +400 ppB step change

The model parameters developed in fitting the above data were applied to two other cases. In Figure 7, the model

predictions for the effect of sample line length on the wet-up of an electropolished and SilcoNerted sample line

are shown. The same adsorption/ desorption parameters are used as were used to approximate the 1 ppm data

shown previously. The only parameters changed in the model were the line lengths, the flow rate (now 10 slpm)

and the inlet concentration (now going from zero to 100 ppb). In Figure 8, we extend the analysis to include an

examination of the effect of a step change from 100 ppb to 500 ppb moisture, with other conditions as in Figure 7.

Such experiments would be difficult, expensive and time-consuming to perform in the laboratory. Assuming the

model is correct, it shows that a 200 meter EPS sample will equilibrate to a 100 ppb wet-up in one hour and that it

will respond to a step change to 500 ppb and achieves equilibrium in under 40 minutes. Both of these results are

theoretical and require laboratory confirmation.

Page | 96

While the model has been developed based on water vapor adsorption desorption characteristics, the same

general trends are seen for trace sulfur species such as H2S and methyl mercaptan, as well as for species like

ammonia.

Testing performed at the Shell Research center demonstrated that conventional stainless steel tubing can adsorb

a significant quantity of mercaptans and greatly delay response times in comparison to electropolished and

sulfinerted tubing 28.

Figure 9. Sulfinert® treated tubing (red) does not adsorb methyl mercaptan (500ppbv), giving accurate

results with no delay.

Similar testing was performed to determine impact of sample line materials on response time to changes in H2S

concentrations. Again the trend is clearly seen in that the electropolished and sulfinerted sample lines

demonstrate faster response speeds and quicker stabilization times as compared to untreated or less treated

lines.

Page | 97

Figure 10. Effects of tubing material of construction on response time.

In a recent paper, Adsorption of dynamically diluted ammonia at part-per-billion to low part-per-million

concentrations in dry nitrogen was studied with treated and non-treated stainless steel and polymer test tubes.

The treatments included electropolishing and two types of coatings based on amorphous silicon29. In this very

thorough work, the authors were able to quantify the number of adsorption sites per square centimeter of tubing

surface, which determines the adsorption capacity of the line. The results clearly show that the combination of

electro-polishing and sulfinerting the sample lines reduces the adsorption capacity and number of surface sites by

more than a factor of 20 relative to a conventional stainless steel line.

Tubing Material Number of Sites

EP with Sulfiner 4.7 x 1012

Sulfinert 14.6 x 1012

EP SS316L 72 x 1012

SS316L 138 x 1012

The authors have as yet been unable to find any data on the adsorption desorption characteristics of heavier hydrocarbons (hexanes, heptanes, octanes and nonanes) on stainless steel surfaces. However, given that these species are present in natural gas applications and have bearing on both the energy content and the hydrocarbon dewpoint of the gas, further investigation is warranted.

-0.5

0

0.5

1

1.5

2

2.5

3

-9.7

-4.0

1.7

7.3

13.0

18.7

24.3

30.0

35.7

41.3

47.0

52.7

58.3

64.0

69.7

75.3

81.0

86.7

92.3

98.0

103.7

109.3

115.0

120.7

126.3

132.0

137.7

143.3

149.0

Time (min.)

Concentr

ation (ppm

)

H2S Sample

Commercial Fused Silica

TrueTube FS - Chemically Polished withSilcosteel

TrueTube EPS - Electropolished withSulfinert

Page | 98

RECOMMENDATIONS FOR INSTALLATIONS

While it is always difficult to establish installation guidelines that fit every application, the following considerations

should be taken when evaluating the design of a sample system, especially when it is to be used for the timely

and accurate measurement of trace components.

Sample Line Materials

All the data presented are consistent with the fact that the use of electropolished and sulfinerted sample lines

reduces the adsorption of reactive species such as hydrogen sulfide (H2S), water vapor (H20), Mercaptans

(CH3SH) and ammonia (NH3). While utilization of such materials increases the initial capital cost of an installation,

the impact is relatively small in terms of overall project scope. The benefits can be dramatic, such as a 20 fold

reduction in surface are for adsorption and dramatically improved response speeds and analytical accuracy.

Heat Tracing

The use of heat traced lines (where possible) is strongly recommended. As shown in Figure 11, the rate of

desorption doubles for about every 50 °F the sample line temperature is increased. Increasing the desorption

rate reduces the total amount of material the sample line can adsorb at equilibrium and increases the response

speed.

Figure 11. Effect of Temperature on the rate of desorption from surfaces (assumed activation energy of

22 KJ/mole)

An additional benefit is that heat traced lines prevent the daily or diurnal temperature changes which can occur

and result in inaccurate readings during the temperature change.

In the event that heat tracing is impossible or impractical, it becomes even more important to consider the effects

of sample line materials, lengths and diameters. Materials should definitely be chosen to minimize adsorption,

and it will be beneficial to insulate the line. The insulation will not prevent the line from experiencing temperature

changes due to ambient conditions, but will slow down the rate at which those temperature changes occur,

allowing more time for equilibration and reducing the large concentration changes which may occur during

ambient temperature swings.

Sample Line Lengths

Sample line runs should be made as short as possible. Each time the sample line is doubled, the number of

adsorption sites is doubled as well so the system will take longer to come to equilibrium. In addition, if the flow

0

2

4

6

8

10

12

14

16

0 50 100 150 200 250

rela

tiv

e D

eso

prt

ion

Ra

te

Temperature ◦F

5025 75 100◦C

Page | 99

rate remains the same, the residence time in the line doubles (giving the gas more time to react) and the first

order lag doubles ( the time it would take for the gas to transit the tube even if there was no adsorption).

Sample line diameter

Reducing the sample line diameter has two effects, it increases the gas velocity (for the same flow rate) and it

reduces the surface area of the tube. Increasing the gas velocity will result in the gas spending less time in the

sample tube and having less time to react. Reducing the surface area reduces the overall number of adsorption

sites and thereby also improves response time and system bias due to adsorption. The combination is a very

effective means to ensure rapid response. Caution has to be taken however. Narrow bore sample lines (such as

1/8” OD tubing) should only be used on streams which will be clean and free of major particulate or condensable

material. Fortunately in natural gas applications this is usually the case. Care must also be taken to ensure that

the pressure drop along the length of the line is calculated and is acceptable. Such calculations are presented in

Reference 2.

SUMMARY AND CONCLUSIONS

Critical factors which affect adsorption and desorption processes in sample lines and gas distribution systems

include the surface roughness, surface chemistry, pressure and temperature. The surface chemistry and

temperature strongly affect the desorption rate, and therefore impact system response speed. Rather than solve

the complex partial differential equations shown previously, a simplified set of finite difference equations has been

presented as a means to model the mass transport problem.

Some comparison of the model to empirical data has been performed, and it appears the model is consistent with

the general trends seen in empirical results. Additional work is required to address deficiencies in the model (such

as the inclusion of only one type of adsorption site), but the model already appears useful as a means of

predicting experimental results and allowing for rapid characterization of the effects of changing process

variables.

The use of the model results has allowed for some general recommendations in regards to sample line

installations.

REFERENCES

1. Donald P. Mayeux, Use of Equations of State (EOS) software, American School of Gas Measurement Technology, 2010

2. Tony Waters, Industrial Sampling Systems: Reliable Design and Maintenance for Process

Analyzers, published by Swagelok Company, 2013

3. David J. Fish, Practical Considerations Of Gas Sampling Systems, Pipeline and Gas Journal, Vol. 239

No. 7, July 2012

4. Donald P. Mayeux, Advances in Natural Gas Sampling Technology, American School of Gas

Measurement Technology, 2002

5. API Manual of Petroleum Measurement Standards , Chapter 14 – Natural Gas Fluids Measurement,

Section 1, - Collecting and Handling of Natural Gas Samples for Custody Transfer, American Petroleum

Institute, Washington D.C. February 2006

6. EEMUA 128:1988, Design and Installation of On-Line Analyzer Systems, Engineering, Equipment and

Materials Users Association, London, United Kingdom

7. Shannon Bromley, Sampling Wet Natural Gas For BTU and Moisture Analysis, NGSTech 2011, New

Orleans, LA , 2011

Page | 100

8. Donald P. Mayeux, Sampling and Conditioning of Natural Gas Containing entrained liquids,

http://www.afms.org/Docs/sample/Sample_and_Conditioning.pdf

9. Tony Waters, Industrial Sampling Systems: Reliable Design and Maintenance for Process

Analyzers, published by Swagelok Company, 2013 pg. 47.

10. Silvia Martinez and Jan Pijpelink, Protect Natural Gas Sample Integrity and Prevent Sulfur Loss with

Sulfinert® Sample Cylinders, http://www.restek.com/pdfs/PCAN1290A-UNV.pdf

11. Benesch, R. Haouchine,M. and Jacksier, T. “The Preparation of Low Concentration Hydrogen Sulfide

Standards”, Gulf Coast Conference, 2002.

12. Lowry, P. and Roll D., “Comparing the Characteristics of surface passivated and electro polished 316L stainless steel”, Report, www.astropak.com.

13. Ohmi T., Nakagawa Y., Masakazu N., “Formation of Chromium Oxide on 316L Austenitic Stainless Steel”, Journal of Vacuum Science Technology, A14(4), 1996.

14. Walls, M.G., et al, “In Situ Observation of the Oxidation and Reduction Processes in Fe-Cr Alloys”, Journal of Vacuum Science Technology, 1996.

15. Joly, J.P., “Temperature-programmed desorption (TPD) of water from iron, chromium, nickel and 304L stainless steel”, Vacuum 59, 2000.

16. Benesch R., Haouchine M., and Jacksier T., “The Stability of 100 ppb Hydrogen Sulfide Standards”, Anal. Chem., 2004.

17. Banes, P.H., “Passivation: Understanding and Performing Procedures on Austenitic Stainless Steel Systems” Pharmaceutical Engineering, Nov. /Dec. 1990.

18. Tuthill, A.H., “Stainless Steel: Surface Cleanliness”, Pharmaceutical Engineering, Vol 14(6) 1994. 19. Gonzalez, M.M., “Stainless Steel Tubing in the BioTechnology Industry”, BioTechology/Pharmaceutical

Facilities Design, 2001. 20. http://www.silcotek.com/sites/default/files/pdfs/silicon-coatings-specifications-summary.pdf, 21. US Patent #6,444,326, 22. Chun I., Cho B., Chung S., “Outgassing rate characteristic of a stainless‐steel extreme high vacuum

system”, J. Vac. Sci. Tech. 1996. 23. Behr P. Terziyski, A., Zellner R., “Reversible Gas Adsorption in Coated Tube Wall Reactors”, Z. Phys.

Chem., 2004. 24. http://www.delstar.com/electropolishing/characteristics-of-the-electropolishing-process.htm 25. Dheanddhanoo, S., Yang J., Wagner M., “Modelling the Characteristics of Gas System Drydown”, Solid

State Technology, 2001. 26. Harris, P., “Relative Response Time of TrueTubeTM when Measuring Moisture Content in a Sample

Stream, HariTec Scientific & Engineering Support, May, 2004 27. Barone, G., Smith, D., Higgins, M., Rowan, S., Gross, W., Harris, P., “Impact of Sampling and Transfer

Component Surface Roughness and Composition on the Analysis of Low-Level Sulfur and Mercury Containing Streams”, Restek Corp., O’Brien Corp., Haritec LLC, ISA Symposium, October, 1995.

28. Application of TrueTube™ in Analytical Measurement Cardinal UHP August 2004. The authors thank the staff at Shell Research and Technology Centre, Amsterdam, for data used in evaluating sulfur gas uptake and memory effects of tubing substrates.

29. O. Vaittinen, et al, Adsorption of ammonia on treated stainless steel and polymer surfaces, Applied Physics B, 2013

Page | 101

NGL SAMPLING

Eric Estrada, Targa Resources

Introduction

The purpose of this paper is to provide the user with a brief overview of what is required for NGL (Natural Gas Liquid) measurement and the importance of proper sampling and analysis.

NGL Measurement

NGL’s are hydrocarbons liquefied by gas processing plants containing ethane, propane, butane, and natural gasoline. Because of the varying size of molecules in an NGL stream, NGL’s are susceptible to solution mixing effects. Solution mixing occurs when compounds containing different sized molecules are mixed together resulting in the smaller molecules fitting into the voids created by the structure of the larger molecules. A simple example of this effect can be described using sand and gravel. Assume a person is building a 4” thick patio base consisting of 2” of sand and 2” of gravel. In calculating the volume of each constituent required to produce a 4” high base, one would assume that they would simply be require equal amounts of sand and gravel. When the base is built up, there is astonishment when the resulting base is much less that 4” high with a large majority of the sand settled within the gravel. Similarly, when hydrocarbons of varying sizes are mixed together, volumetrically there will always be some solution mixing effect resulting from smaller molecules mixing in between larger molecules. Figure 1 shows molecular models of Normal Butane, Propane, and Ethane respectively and the variability in the size of the molecules.

Figure 1 - Molecular Models of N-Butane, Propane, and Ethane Respectively

Because of solution mixing, it is difficult if not impossible to create volumetric correction algorithms that compensate for pressure and temperature effects on NGL’s due to the near infinitely possible combinations of hydrocarbons in an NGL stream. For this reason, mass measurement is the preferred means of measuring NGL’s since mass is constant and is either destroyed or created during the measurement process.

Mass measurement involves the determination of the mass through a meter either directly or indirectly, and the determination of the composition of the NGL stream. Since the main products within an NGL stream which include, ethane, propane, iso-butane, normal butane, and natural gasoline, are priced volumetrically, conversion from mass to volume is a required step in the process which is where the composition of the stream becomes important.

Figure 2 shows a mass calculation example involving indirect mass calculation. The mass is derived as follows:

Mass=IndicatedVolume×MeterFactor×FlowingDensity

Where the units of measure for the following example are:

IndicatedVolume - Barrels

FlowingDensity - lb/barrel

The flowing density is converted to pounds per barrel as follows:

FlowingDensity (lb/barrel) = FlowingDensity (g/cc) * 8.345406 (g/cc per lb/gallon) * 42 (gallons/barrel)

So the resulting calculation is:

Mass = 10000 0.9907 0.4878 8.345406 42 = 1693873× × × ×

Page | 102

Once the mass is determined, the composition from a representative sample is then used the mass of each component by multiplying the mass by the weight fraction of each component. It should be noted that if the analysis is in mole or liquid volume percent, it will need to be converted using the molecular weights and/or liquid densities found in GPA 2145 or other standard as dictated by contract.

The hexanes plus density and/or molecular weight is generally determined by performing an extended analysis of the sample to thoroughly characterize the hexanes plus portion of the stream.

NGL Sampling

Before discussing the requirements and steps involved in proper sampling and analysis of NGLs, it is important to understand the impact that the NGL analysis has on the value of the NGL stream. In Figure 3, the starting composition is from an actual NGL analysis and using January, 2014 pricing, the resulting value of this product in $38.67 per barrel. The sample cylinder was subsequently opened and the pressure reduced by 100 psig. The net effect was a loss in the light ends of the sample and thus the value of the product increased to $39.92 per barrel. Over a period of a year, assuming that this stream flows 50,000 barrels per day, the error would result in a financial impact of 22 million dollars. So proper sampling and analysis is extremely crucial to NGL measurement and can result in large financial losses and gains.

Gross 10000 Barrels Notes: 1) From GPA 2145

Meter Factor 0.9907

Flowing Density 0.48780 g/cc

Mass* 1693873 Lbs

* Mass = Gross bbls* Meter Factor*Density (lb/bbl)

Density (lb/bbl)=Density (g/cc)/0.1198264 (g/cc)/(lb/gal)*42 (gal/bbl)

(A) (B) (C) (D) (E)

Component

Lbs/Gallon

(1)

Weight

Fraction Mass

Component

Mass

Barrels @ 60F

and EVP

B * C E / A

N2 6.727 0.00000 1693873 0 0

CO2 6.8534 0.00054 1693873 906 3

C1 2.5 0.00384 1693873 6503 62

C2 2.9716 0.35662 1693873 604076 4840

C3 4.2301 0.30772 1693873 521241 2934

IC4 4.6934 0.05778 1693873 97871 496

NC4 4.8696 0.13739 1693873 232717 1138

IC5 5.2074 0.03890 1693873 65888 301

NC5 5.2618 0.03684 1693873 62401 282

C6+ 5.5363 0.06038 1693873 102268 440

1.00000 1693873 10497

Figure 2 – Mass Conversion to Component Volume

Page | 103

So what are the proper steps to take when sampling NGLs?

• Proper System Design

o A representative sample is necessary to properly characterize the fluid. This involves either extracting a small portion of material from a bulk tank that represents the total of the bulk tank, or extracting small samples over a pipeline movement. In order to be a representative pipeline sample, the samples must be obtained proportional to flow throughout the batch. A time proportional sample will lead to compositional errors since NGL composition will vary whether the NGL’s come from a gas plant or from storage.

o When sampling from a pipeline, sample probes must be used in order to obtain a representative sample. The sample probe should be located in the center third of the pipe and should be 3 to 10 diameters from piping elements and power mixers, or 0.5 – 4 diameters from static mixers. The preferred location for sample probes is on horizontal pipe, but they can also be installed on vertical pipe as long as they are on the down flow section of the pipe.

o The sample source should always be homogeneous, i.e. a single phase liquid.

• Proper Sample Handling

o Samples can be obtained through a variety of techniques as described in GPA 2174. The most comment method for obtaining a sample is through the use of a gloating piston cylinder. While the equipment involved is more expensive than using single cavity cylinders, the advantages include maintaining a constant pressure on the sample which means no loss of light ends during the sampling process.

o The use of a floating piston cylinder requires and inert gas on the back side of the piston which should be 100 psi over the process pressure. The cylinder is filled by slowly connecting it to the process line of composited sample vessel and slowly opening the inter valve thus drawing in the liquid sample.

o In order to prevent over pressure of the cylinder through thermal expansion, the cylinder should be filled to no more that 80% of its capacity.

• Proper Analysis Determination

o While proper sampling technique is important to the sampling process, another key step in the process is using gas chromatography. Two standards are fundamental to proper analysis of an NGL sample. GPA 2177 describes the method used to analyze NGL samples using gas chromatography and also specifies the reproducibility and repeatability requirements for gas

Starting LV%

January 2014

$/Barrel Ending LV%

N2 0.000 $ - 0.000

CO2 0.000 $ - 0.000

C1 0.678 $ 11.64 0.473

C2 45.186 $ 11.64 42.948

C3 27.261 $ 52.76 28.116

IC4 4.790 $ 60.35 5.207

NC4 11.510 $ 59.04 12.220

IC5 2.912 $ 89.61 3.172

NC5 2.748 $ 86.91 2.863

C6= 4.910 $ 86.91 4.996

Average $38.67 $39.92

Value @ 50,000 BPD $1,933,638 $1,995,900

Difference per Year $22,725,838

Figure 3 - NGL Pricing Example

Page | 104

chromatographs. GPA 2186 describes the method for determining the hexanes plus composition and properties on a liquid hydrocarbon sample.

Common Pitfalls Encountered In NGL Sampling and Measurement

Proper determination of the various parameters used for mass measurement is crucial to minimizing errors. First and foremost is the proper determination of the composition during the product movement. In Figure 4, the same mass is used as in Figure 2, but the ethane is reduced by 0.1% and the propane increased by 0.1%. The result is a 3-barrel difference, or 0.02% between the two calculations.

Similarly, the characterization of the hexanes plus portion of the stream is also critical in determining the proper volume. In Figure 5, the hexanes plus density has been changed from 5.5363 in Figure 2 to 5.8414 or increased by 5.5%. The calculation yields 10,462 barrels, which is 35 barrels or 0.30% less than the calculation in Figure 2. Incidentally, the hexanes plus density in Figure 5 was from an actual extended analysis while the 5.5365 value is density of Normal Hexane. Another error is to use the

density/specific gravity of the analysis as the flowing density in the calculation of mass. Figure 6 uses the gravity from the analysis to calculate the mass. The result is a volume of 9,943 barrels which is 554 barrels or 5.28% less than the volume calculated in Figure 2.

Lastly, the other common error that is made when performing mass calculations is the use of the wrong standard densities. In many cases this error is small, but should be investigated when verifying the integrity of a mass calculation.

Conclusion

Due to the nature of NGL’s and their propensity to experience solution-mixing effects, it is desirable to determine the volume of product using mass measurement techniques. A key component to mass measurement involves proper determination of the NGL composition by both proper sampling and analysis. Errors in either sampling or analysis can lead to significant errors in the value of GL’s.

References

GPA 2174 – Obtaining Liquid Hydrocarbon Samples for Analysis by Gas Chromatography

GPA 2177 – Analysis of Natural Gas Liquid Mixtures Containing Nitrogen and Carbon Dioxide by Gas Chromatography

Gross 10000 Barrels

Meter Factor 0.9907

Flowing Density 0.48780 g/cc

Mass* 1693873 Lbs

(A) (B) (G)

Component LV%

Lbs/Gal

lon (1)

Barrels @

60F and

EVP

F / B

N2 0.00% 6.727 0

CO2 0.03% 6.8534 3

C1 0.59% 2.5 62

C2 46.11% 2.9716 4824

C3 27.95% 4.2301 2924

IC4 4.73% 4.6934 495

NC4 10.84% 4.8696 1134

IC5 2.87% 5.2074 300

NC5 2.69% 5.2618 281

C6+ 4.19% 5.8418 438

100.00% 10462

Gross 10000 Barrels

Meter Factor 0.9907

Flowing Density 0.46205 g/cc

Mass* 1604456 Lbs

(A) (B) (G)

Component LV%

Lbs/Gal

lon (1)

Barrels @

60F and

EVP

F / B

N2 0.00% 6.727 0

CO2 0.03% 6.8534 3

C1 0.59% 2.5 59

C2 46.11% 2.9716 4585

C3 27.95% 4.2301 2779

IC4 4.73% 4.6934 470

NC4 10.84% 4.8696 1078

IC5 2.87% 5.2074 285

NC5 2.69% 5.2618 267

C6+ 4.19% 5.5363 417

100.00% 9943

Gross 10000 Barrels

Meter Factor 0.9907

Flowing Density 0.48780 g/cc

Mass* 1693873 Lbs

(A) (B) (G)

Component LV%

Lbs/Gal

lon (1)

Barrels @

60F and

EVP

F / B

N2 0.00% 6.727 0

CO2 0.03% 6.8534 3

C1 0.59% 2.5 62

C2 46.21% 2.9716 4852

C3 27.85% 4.2301 2924

IC4 4.73% 4.6934 497

NC4 10.84% 4.8696 1138

IC5 2.87% 5.2074 301

NC5 2.69% 5.2618 282

C6+ 4.19% 5.5363 440

100.00% 10500

Figure 2 - Ethane and Propane

Adjusted Figure 5 - Changed C6+ Gravity Figure 6 - Using Standard

Density in Place of Flowing

Density

Page | 105

GPA 2186 – Method for the Extended Analysis of Hydrocarbon Liquid Mixtures Containing Nitrogen and Carbon Dioxide by Temperature Programmed Gas Chromatography

GPA 8173 – Method for Converting Mass of Natural Gas Liquids and Vapors to Equivalent Liquid Volumes

GPA 8182 – Standard for Mass Measurement of Natural Gas Liquids

Page | 106

Injection of Chemicals and Their Impact on Sampling Jay St Amant – A+ Corporation

Brad Massey - Williams Introduction Historically natural gas producers have had the necessity to chemically treat gas streams for suitability of their intended purpose. One of the most common is methanol injection while others include corrosion and sulfur inhibitors. Each chemical has its own properties and manner in which it interacts with the natural gas stream and has to be dealt with accordingly. Methods and chemical types are ever changing as the evolution of natural gas production and gathering brings new challenges.

Shale Gas Compositions The production and transport of shale gas is posing some unique challenges for the natural gas industry. The existing infrastructure for producing and transporting “conventional” natural gas, which has been the norm for nearly the past 100 years, has been optimized to efficiently process and deliver gas that meets end-user requirements for heating value, hydrocarbon dew point, and contaminant content. Produced shale gases observed to date have shown a broad variation in compositional makeup, with some having wider component ranges, a wider span of minimum and maximum heating values, and higher levels of water vapor and other substances than pipeline tariffs or purchase contracts typically allow.

Significant differences in levels of ethane, propane, hexane and heavier components, and diluents (primarily CO-2 and nitrogen) have been seen among the various shale formations. These, in turn, result in significant differences in the heating value, Wobbe number, and other parameters that guide end-use applications of natural gas. Because of these variations in gas composition, each shale gas formation can have unique processing requirements for the produced shale gas to be marketable. Ethane can be removed by cryogenic extraction while carbon dioxide can be removed through a scrubbing process. The typical maximum recommended amount of total inert components (CO-2, nitrogen, etc.) is 4 mole %. When shale gas is out-of-tolerance per the FERC provisions, the options available to the gas producer and pipeline operator are the same as the options for dealing with other gas supplies with high levels of unacceptable components. Restricting unacceptable gas from entering the transmission pipeline grid is always an option, but is often the choice of last resort, since replacement gas supplies may be hard to obtain. Processing of the gas – extraction of hydrocarbon condensates, dehydration to remove water, and “scrubbing” to remove CO-2 and hydrogen sulfide – is a common approach, but may not be pursued if it is not cost-effective.

Shale Gas (lower 48 U.S.)

Page | 107

Monitoring Shale Gas Quality Diluents and contaminants, such as water vapor, oxygen, and hydrogen sulfide, may be found in shale gas and thus must be monitored using analytical equipment designed for the intended application. The measurement accuracy of these devices can vary greatly, depending on the technology employed. For example, length-of-stain tubes and laser-based analyzers are both technologies that measure moisture content of a natural gas stream. Length-of-stain tube kits can be purchased for a few hundred dollars, but current standards for their use cite an accuracy of ±25% of reading. A single laser-based analyzer costs roughly $15 to $20 thousand dollars but manufacturer-stated accuracy is typically ±1% of reading. For this reason most operators only use the length-of-stain tube results as an initial indicator and use the more reliable higher accurate technologies of electronic analyzers for tariff enforcement. Selection of the best equipment for monitoring a particular shale gas stream should take into account not only the initial capital cost, but also operational and maintenance costs, plus, the potential costs associated with inaccurate diluent and contaminant measurements - including additional gas-processing costs, costs associated with pipe corrosion or other equipment damage, and expenses associated with unanticipated shut-ins. Contaminants, such as CO-2, hydrogen sulfide, and water vapor, must stay within tariff limits for a transmission pipeline or local gas distribution company to accept a gas stream and errors in gas volume or heating value of even a fraction of a percent can be very costly over time.

H2S Scavenger Application Hydrogen sulfide gas (H2S) is dangerous at certain levels and very corrosive to well and surface equipment, thereby requiring natural gas tariff specifications to be typically restricted to 4 ppm. H2S gas is also present in sour crude, produced water and condensates, and can be a hazard in the headspace of tanks that contain these liquids. H2S gas scavengers remove H2S from process systems to improve safety, protect equipment and meet tariff specifications.

Continuous injection or batch treatment using scavengers are both used as effective removal methods, however continuous injection is more commonly used. Batch treatment bubbles sour gas through the solution contained in a tower where as the injected method uses a pump and an atomizer to disperse the chemical directly into the gas stream. The rate of use ranges from 1-to-200 pounds per day to effectively remove H2S. The amount required and its effectiveness is affected by several factors, including gas temperature, velocity, amount of liquids in the line, retention time available to remove the H2S and the ability to adequately inject the scavenger into the gas stream. With good temperature and atomization, H2S can often be removed from the system in as little as 15 seconds.

There are a variety of formulations for complete and partial H2S removal in both oil-soluble and water-soluble applications. Due to the nature of some H2S scavenger chemistries, a co-product of the scavenging reaction is an amine component that can raise the pH of water in the produced gas stream. If a calcium carbonate scaling potential exists in the water, it could be enhanced by the increased pH of the fluid. To prevent the occurrence of unwanted scale formation, a specifically formulated additive package is used to control calcium carbonate scale in severely scaling systems where an H2S scavenger is required.

Page | 108

Below is an evaluation site where some H2S was present and Scavengers are being injected.

Date: March 14, 2013. Technician Contact Information: James Tabor James.tabor@kinderhawk.com Location: Kinder Morgan-Kinderhawk Field Services-Coushatta Plant Inlet in Haynesville Shale Area (near Marksville, LA) Information given by technicians about test location when on site:

Operating pressure 1066 PSI

Temperature of sample is 62-100 degrees F

40#s moisture according to stain tube

Probe feeding two SpectraSensors analyzers- an H2S & a CO2

o 6 L/min total flow through the probe

Other notes about site:

Probe Regulator installed and would not regulate for more than 1-2 days

They are injecting sulfagaurd on location just a few feet downstream of the sample point

Laser based H2O and CO2 Monitors

Page | 109

Corrosion Inhibitor Tank and Label

Membrane Tipped Regulator Probe

Obtaining an analytically correct sample begins with selecting a proper sample point. In the above application, the sample point is within a few feet of the chemical injection point. The probe has a 4” insertion depth into the meter section of the pipeline. The tip of the probe is at times submerged in the liquid chemical. It will not be possible to obtain an analytically correct sample at this sample point. In some cases the operator is not looking for an analytically correct sample but rather is looking for trending. So in this application, the operator would not move the sample point. To continue to use this sample point, the gas sample probe is removed and a liquid sample probe is used with an external filter/heated regulator.

Page | 110

Methanol Injection Methanol is often times injected into a gas stream as a dehydration method to prevent freezes and hydrate formation. Hydrate inhibition occurs in the aqueous liquid, rather than in the bulk vapor or oil/condensate. While most of the methanol dissolves in the water, a significant amount of methanol either remains with the vapor or dissolves into the liquid hydrocarbon phase. Even though the concentration of methanol in the vapor or liquid hydrocarbon is small, with low water amounts, the majority of methanol may be consumed by the vapor or liquid hydrocarbons because the hydrocarbon-phase fractions are much larger than the water-phase fraction.

Although methanol injection is often necessary, careful consideration of the location relationship between the sample extraction point and methanol injection point should be made. This will reduce the risk of contaminant migrating into the sample system.

Corrosion Inhibitors Corrosion inhibitors are used in a variety of ways such as during pipeline construction they are combined with water during hydrostatic testing of the pipeline. They can also be injected throughout the service life of some systems. The problem with water soluble inhibitors is that their retention and adhesion to pipe surfaces can’t be obtained without the addition of binding agents. Below are some additive comments from the patent filing for one corrosion inhibitor. Note that the actual compound only contains about 10% corrosion inhibitor and the remaining 90% are binding agents.

12.A method according to claim 9, wherein said binding agent component comprises one or more

of the following compounds: polyacrylamide, fish glue, highly-substituted hydroxypropyl guar (MS

1,2), monoethylene glycol.

13. A method according to claim 9, wherein said water-soluble corrosion inhibitor composition

comprises around 90% binding agent and around 10% corrosion inhibitor component.

14. A method according to claim 12, wherein said binding agent component is polyacrylamide

(250,000-1,000,000 Mw polymer).

15. A method according to claim 9, wherein said binding agent component contains fish glue and

polyvinyl alcohol.

Many corrosion inhibitors contain other additives to enhance the performance of their products. A list from one manufacturer shows several additive options.

Oil wetting and solids dispersing agents Corrosion inhibitors Scale inhibitors Anti-foulants Oxygen scavengers Foaming agents.

Also as seen below in the same patent filing as shown above, there may be the following.

[0062] Many additional components may be included in the compositions of the invention, according to the individual requirements of each system. For example, scale inhibitors are sometimes required to be present within a “corrosion inhibitor package”. These types of products use polyacrylates and phosphonate chemistries, for example, phosphonates, acrylic co/ter-polymers, polyacrylic acid (PAA), phosphino carboxylic acid (PPCA), phosphate esters, or other traditional aqueous-based scale inhibitor chemistries. The concentration of scale inhibitor as incorporated in the compositions of the invention is in the order of about 1-5% by weight.

Page | 111

[0063] Demulsifiers may also be included within the compositions of the invention. These products are propylene oxide/ethylene oxide co-polymers and resin formulations. They are hydrophilic molecules that attach onto water molecules, which are emulsified in oil, and cause the water to sink to the bottom. These products may be incorporated into the composition in the event that there should be concern about the oil production emulsifying the water residue upon start-up, causing a top-side process problem. [0064] Anti-foams may also be incorporated into the compositions of the invention, as some corrosion inhibitors and biocides tend to stabilise foams, due to their surfactant nature. Anti-foams eliminate this tendency. Such compounds are generally polyglycol-based chemistries, and should be present in proportions of around 0.5% wet weight of the composition. [0065] Wax inhibitors may also be added to the compositions of the invention, as this helps to prevent the build up of wax on the pipe wall. This is a particular problem for certain crude oil types. Wax inhibitors are polymeric-based and generally incorporate an n-alkane backbone and can incorporate PEG ester groups. If included, wax inhibitors should be present in a proportion of around 5-10% wet weight of the composition. Fluids and Additives used in Hydraulic Well Fracturing

Page | 112

It is important to note: Not all of the additives are used in every hydraulically fractured well; the exact “blend” and proportions of additives will vary based on the site-specific depth, thickness and other characteristics of the target formation. Biocides often exception because of long-term importance to “health” of reservoir.

Source: “WATER-RELATED ISSUES ASSOCIATED WITH GAS PRODUCTION IN THE MARCELLUS SHALE. URS Corporation. Fort Washington, PA. May 21, 2010. Water Consulting Services in Support of the Supplemental Generic Environmental Impact Statement for Natural Gas Production, NYSERDA Contract PO Number 10666. Pg. 2-4).

Page | 113

Applications of Biocides

The presence of bacteria and other microorganisms in deposits, transmission systems and in stored products is a very unfavorable phenomenon and often difficult to remove. A necessary condition for the emergence and development of micro-organisms is the presence of water and a carbon source in a given environment (fuel tanks, pipelines, gas storage, gas supply systems, etc.). The issue of deterioration of natural gas resulting from the metabolic activity of microorganisms is related to two main aspects – gas transmission and storage of this raw material in underground gas storage (UGS), i.e. exploited geological structures. Corrosion of metals, in this particular case of gas pipelines is one of the major problems of the gas industry, causing enormous economic losses. It is estimated that 40% of the corrosion occurring in the interior of gas pipelines is caused by the action of microorganisms (MIC – microbiologically influenced corrosion), mainly sulfate-reducing bacteria. Attempts to eliminate microorganisms involve using chemicals, exhibiting biocidal properties, which besides the physical method is the most popular and most effective technique of eliminating microbiological contamination. The selection of appropriate antibacterial or antifungal agents requires the consideration of factors affecting the efficiency of the process. Such measures are primarily sought, which show the greatest spectrum of activity. Given the mechanism of chemical action of biocides, biocides can be divided into two groups, i.e., substances with oxidizing and non-oxidizing effect. The most commonly used oxidizing biocides are chlorine, bromine, ozone and hydrogen peroxide. However, use of oxidizing biocides is associated with negative effects:

• interaction with other chemicals (corrosion inhibitors), • the possibility of interaction with non-metallic substances, • initiation of corrosion of structural materials.

Before each treatment with oxidizing preparations, these effects should be taken into account when considering the potential for oxidation, the dose and type of treatment (intermittent or constant). The group of non-oxidizing biocides includes aldehydes (formaldehyde, glutaraldehyde), acrolein, quaternary ammonium compounds, amines and diamines [29], and isothiazolones. Often used in the industry, quaternary ammonium compounds are used as cationic corrosion inhibitors and biocides. Biocidal activity of these substances is to dissolve the lipid cell membrane, which leads to loss of the cell contents of the microorganism. Additionally, QUATS prevent the formation of polysaccharide secretions during bacterial colonization, thus showing antibacterial activity. Often, biocides using QUATS have water, alcohol or potassium base as the dissolving phase, the use of alcohol increases the antibacterial activity of the preparation, since alcohol has biocidal abilities and facilitate penetration of an entity into the cell. QUATS generally work best in an alkaline environment. Inhibition of corrosion using these substances is to create a thin protective layer on the inner parts of the installation, thereby the possibility of interaction of oxidizing agents with steel components of the installation is reduced. Furthermore, these compounds were studied as a control substances, and even as substances preventing biofilm formation. These preparations are used in closed systems and gas manifolds. However, they are not used during the exploitation of oil because they may adversely affect the permeability of the crude oil deposit. Furthermore, they are not compatible with oxidizing agents, especially the chlorates, peroxides, chromates or permanganates. Most of these compounds are readily biodegradable. Another type of

Page | 114

biocides is isothiazolones. They are fast-acting biocides inhibiting growth, metabolism and biofilm formation by algae and bacteria. They are used in combination with other biocides or individually, typically aqueous solutions of chloride- and methyl-derivatives of these compounds are used. Isothiazolones are used only in an alkaline medium, at pH < 7 they lose biocidal properties, moreover, these compounds can be used in combination with other chemicals without changes in performance. An exception is environment containing hydrogen sulfide, which causes deactivation of isothiazolones. The main application areas of isothiazolones are coolants and cooling and lubrication fluids. A compound commonly used in biocidal formulations is methylchloromethyl-isothiazolone (MCMI) and 4,5-dichloro- 2-n-octyl-4-isothiazolin-3-one (DCOI), which causes inhibition of dehydrogenase and thus interferes with metabolic pathways. The following is a formula of this biocide: In addition, commonly used substances with biocidal action are compounds containing an aldehyde group, which include:

• formaldehyde, • 2-propenal (acrolein), • ortho-phthalic aldehyde (OPA), • three-seven-carbon (C3-C7) compounds having aldehyde groups (e.g., pentane-1,5-dial)

Application Methods

Not open systems

Chemical additives metered in separately through a closed system.

Dedicated frac chemical injection trucks are deployed, treating all frac fluids at the appropriate

biocide concentration.

Biocides often added between the blender vat “on the fly” and the booster pump on the blender

Low pressure side application

Minimize oxygen entry

Enables safe, as-needed, injection of the biocide and other production chemicals, such as oxygen

scavengers and scale inhibitors, etc. into the frac fluid.

Conclusion Needless to say that none of the aforementioned chemicals are desired to be ingested into a gas sample. The best way to avoid chemical interference and sample distortion is to be aware of their presence and avoid sampling in close proximity to where any chemical injection points are located. When designin a sample system be sure to consult the various suppliers of sample conditioning equipment and make them aware of any chemicals being injected into the system and identify location of the injection point. Each sampling application will have unique conditioning requirements and should be evaluated independently to ensure proper design.

References July 2011, Vol. 238 No. 7 Pipeline & Gas Journal Shale Gas Measurement And Associated Issues By Dr. Darin L. George and Edgar B. Bowles, Jr. | July 2011, Vol. 238 No. 7 Darin L. George, Ph.D., holds the post of Senior Research Engineer, Fluid Dynamics and Multiphase Flow Section, Mechanical Engineering Division of Southwest Research Institute, San Antonio, TX. Edgar B. Bowles, Jr., M.S., holds the post of Director, Fluids and Machinery Engineering Department, Mechanical Engineering Division of Southwest Research Institute, San Antonio, TX. Ph: 210-522-2086, e-mail: Ed.bowles@swri.org. Anna Turkiewicz, Joanna Brzeszcz, Piotr Kapusta Oil & Gas Institute, Krakow “The application of biocides in the oil and gas industry” July 2013 “Water Related Issues Associated with Gas Production in the Marcellus Shale” URS Corporation

Page | 115