Low Risk Flow Measurement Fluids

Low risk fluids:

Considering these characteristics and answering related concerns minimizes the influence of fluids on measurement accuracy, and simplifies metering.

High Risk Flow Measurement Fluids

Many fluids classified as “high risk to measure” become the main consideration in the choice of a meter and in determining the potential for measurement accuracy.

High risk fluids include:

Specific meters may react differently to the problems listed above, and there may be one that works better than others for the specific problem presented. It should be recognized that the fluid sometimes must be measured even if it is a “high risk” fluid, and the cost of making it a “low risk” fluid is prohibitive in a cost/value study. The preferred fluid conditions are sometimes simply not available at the point of measurement (Figure 6-1). On the other hand, these characteristics may result in a measurement uncertainty that is no better than ± 20 to ± 30%. An example is the measurement of carbon dioxide for injection into oil reservoirs for tertiary oil recovery. Removing oil from the formation efficiently requires the CO₂ to be in the “dense phase” stage (near the critical point). Under these conditions—where the temperature and pressure are near the critical points of 88°F and 1,087 psia—the CO₂ density variation may be in the order of one percent per degree Fahrenheit. In this high risk fluid situation, the flow measurement practitioner might prefer to change the temperature, or the pressure, or both, but the successful use of the fluid to remove the oil precludes such changes, so wider limits must be put on this measurement.

Basic Requirements and Assumptions

Fluid flow must be single phase for basic meter accuracy to be meaningful. There are two problems with a two-phase fluid. One is its effect on the meter mechanics, and the other is difficulty of obtaining a truly representative sample to determine the composition for calculating the reduction to base conditions. Studies are under way to determine whether there is a recommended way to sample two-phase fluids. Currently, sampling under these fluid conditions is less than optimal. This is particularly important in oil and gas production operations (Figure 6-2).

For liquids in a gas, the two-phase flow may affect flow profile configurations. Such flow patterns may conform to one of a number of regimes in a pipeline. The first phase tends to be droplets; if they are small enough, they form a homogeneous mixture so both the flow effects and sampling errors for composition mentioned above are minimized and may not be significant. The second regime occurs when there are sufficient droplets for them to begin to accumulate in non-flowing areas, including in the area of the meter, where flow distortion may occur. Also, a sample can be inaccurate. The third regime occurs when additional liquid is present. For example, there may be annular liquid flow with a core flow of gas. In this case, the mechanical configuration and the sampling are in trouble.

The addition of even more liquid means that there will then be two separate flowing streams (usually flowing at different velocities) forming layered flow (Figure 6-3).

The next regime is slug flow, which occurs as liquids collect until the lines are filled to a point where they “burp” the liquids. In such cases, there will be no way to take care of mechanical or sampling problems since liquid slugs will be followed by gaseous pockets with no indication of which fluid is causing the meter to indicate flow. Separation must be done prior to attempting measurement in any of the cases where two-phase flow exists at the meter.

In the reverse case—gas in liquids—similar problems exist, and de-aerators must be used. With solids in liquids, filtering is required to correct measurement and to minimize meter damage.

Some meters that react to mass can be successfully used for two-phase measurement provided they are not used to attempt to calculate volume without additional information on the fluid composition or density.

Some fluids are unstable and may present measurement problems when they break down into other products during interrupted flow conditions, or during exposure to conditions along a line. Hydrates in natural gas measurement are an example. Hydrates, a mechanical combination of hydrocarbons and water that forms an ice-like material, can block off main lines or gauge lines so that readings of differential or static pressure are impossible. Likewise, crude oil can form an emulsion with water, which does not lend itself to flow measurement. Some liquid plastics and such fluids as molten sulfur set up if they are not kept flowing or the meters are not heated. In these cases, the fluids must be removed from the lines when shutdown occurs. The mechanical problems these fluids create may prevent measurement if these precautions are not taken.

Although the problems caused by two phases have been outlined, some users are not aware that condensation of gases often may take place in a meter, since this may be the point with the lowest pressure and temperature in the system. Although the fluid may be above saturation upstream in the system, conditions at the meter may be below the condensing point.

The same is true with flashing liquids; the pressure within the meter may be lower than the pressure existing after flow goes through the meter to “recovery.” If there is any question, a higher back pressure should be planned, rather than risking the fluid flashing (Figure 6-4).

Fluids that are measured near their critical point have variations in relationships which are so large that the measurement of pressures and temperature within tolerances close enough to predict the effect on volume is beyond the ability of measuring equipment. Likewise, the accuracy of PVT correlation data deteriorates as the critical pressures and temperatures are approached. Because of these problems, flow measurement should not be attempted in such areas of operation if this can be avoided. The problem may also be addressed by moving the meter station to a better location in terms of fluid properties.

A number of meters have limitations for lower Reynolds numbers, and such limits should be checked in the standards prior to attempting to use a meter. With new meters the information on these limits is often quite general, and the user may be left in a quandary as to its meaning. If there is any question, a meter should be operated well away from such limits for best accuracy.

In summary: from a measurement standpoint, no fluid should be measured near a point of phase change, fluid characteristics change, near condensation, at too low a Reynolds number, or near critical pressure or temperature. Good flow measurement practice requires that these conditions are recognized and proper precautions are taken to minimize their effects.

Data Sources

Many sources of data for fluid physical properties characteristics are available in the industry. The particular industry standards for specific applications should be the first place to look. There are general fluid specifications available from universities, national standards organizations, and from suppliers that handle certain products; manufacturers of fluid products have their own data. The Flow Measurement Engineering Handbook has accumulated much data that is useful for general flow measurement.

Gas and liquids are usually considered as either “pure products” or “commercial products” in most references. Mixture laws may be used to estimate the combined characteristics of uncommon fluids, but caution should be exercised if the mixtures contain widely varying molecular weights or involve extreme conditions of pressure and/or temperature. In these cases, actual PVT tests should be run over the ranges of operation and a specific set of tables or equations should be set up for the particular mix—if such information can be derived from the results obtained.

It is important to avoid the tendency in industry to specify a mixed fluid only by its major component. Mixture characteristics must be considered, not just the PVT relationships of the pure product. Any data correlation must be examined to determine the fluid data parameters involved before applying the data to a specific metering system (Figure 6-5).

Liquid propane is an example of a liquid that falls into this area of concern. Commercial propane contains 95% propane, while the percentages of the other constituents may vary; this variation will affect the correction factors used. Reagent grade propane, however, allows the use of “pure” correction factors such as those available from the National Institute of Standards and Technology (NIST).

Another way to approach the problem of variable fluid characteristics is to use densitometers or mass meters, and then measure mass. The measured mass can then be converted to an equivalent volume at base conditions using an analysis and the density of the individual components at base conditions. (Note: This approach will be correct to determine a contract volume. But if there are significant differences in the flowing and base volumes, some means of deriving an approximate line condition flow volume may be required for operations.) The contract volume is necessary for changing custody, but the actual volume at line conditions is needed for operational control.

For any petroleum fluid, Chapter 11 of the API Manual of Petroleum Measurement Standards is an excellent data source.

The ASTM, in conjunction with both the API and the GPA, also publishes common standards for petroleum-related fluids. International standards organizations also publish fluid data documents. (See Chapter 4, “Basic Reference Standards” in addition to the list below.)

Gases

The following section discusses problems unique to some commonly measured gases (Figure 6-6).

Natural gas is one of the gases most commonly measured, since it is used both as a fuel and a feedstock in many industries. It represents the largest daily dollar volume of any gas routinely bought and sold. The natural gas industry has been the leader in developing gas measurement technology and its standards are used in related areas of gas measurement.

The ANSI/API Manual of Petroleum Measurement Standards, Chapter 14, “Natural Gas Fluids Measurement,” Section 3, “Concentric, Square Edged Orifice Meters” Parts 1–4 (also entitled “American Gas Association Report Number 3,” Parts 1–4; and “Gas Processors Association 8185” Parts 1–4) is the most common standard used for gas measurement of all kinds. Representing over 70 years of study of gas measurement with the orifice meter, the standard continues to be upgraded further by additional work.

Natural gas as a fluid varies from an easily measured fluid to a very difficult fluid to measure to close limits. Separated, dehydrated, “pipeline quality” natural gas is normally easy to measure since it is a well defined fluid for which very precise data is available for relating the pressure, volume, and temperature from flowing conditions to base conditions. It is normally very clean, with a minimum of solid “pipeline dust” and compressor or dehydration plant oils present. It is normally handled at temperatures and pressures that cause minimal meter design limitations or operating concerns.

Pipelines typically measure the gases within tenths of a percent in their pipeline balances between purchased gas, operating use gas, and gas sold (Figure 6-7).

On the other hand, produced gas is often handled as a saturated fluid (separated to single phase but not dried), and the problems of flow measurement increase. The ability to balance a production field (with multiple wells) is thought to be within the industry norm if this balance is within 3 to 5%.

Other than some operating problems and poor maintenance that may affect measurement, the main cause of error is the fluid characteristics that cause both mechanical problems (liquid in the meter) and errors in fluid density calculation (determining the proper specific gravity or relative density from a sample). Quite often, the volumes measured are not used directly for custody transfer, but for allocation in determining the percent of total volume contributed by each well’s flow.

At the present time, the measurement of two-phase fluid (gas and gas liquids and/or water) is not attempted because of the problems caused in a meter. There is a great deal of work being done on multiphase measurement that may, in time, resolve these problems. Getting a material balance in a system or processing plant can be one of the most frustrating flow measurement jobs. Careful attention must be paid to all of the concerns outlined here; even then, getting a close balance is difficult.

Mixtures of gases are more easily measured if the mixture has a relatively constant composition. This allows specific PVT tests to be run, or data may be available for common mixtures from prior work. The ability to use the mixture laws successfully has been previously discussed. If the mixture is changing rapidly, use of a densitometer or a mass meter may be required to determine an accurate quantity.

Ethane, a common chemical building block, may be a measured product or a mixture of an enriched stream from a processing plant. Data are available on ethane as a pure gas product. Except at critical points, measuring ethane is fairly straightforward. However, ethane’s critical temperature of 90.1°F at a critical pressure of 667.8 psia may easily occur in normal pipeline and process plant operations. For the best measurement, some heating or compression may be required for conditions near critical to make them more favorable for measurement.

From a measurement standpoint, it is worth remembering that though classified as a gas, ethane mixtures have the same characteristics as dense fluids up to approximately 1,000 psia and at temperatures from 90° to 120°F. In this area of operation (which may exist quite often in a pipeline, or as part of a pipeline or processing measurement requirement), the density changes significantly with small changes of temperature and/or pressure. Because of this, flow metering accuracy relates to these and proper selection of the equation of state for the ethane. At pressures below 670 psia, if the temperature drops below 90°F, two-phase flow can be encountered. When either of these situations is likely, consideration should be given to adding heat or pressure to the system prior to attempting flow measurement.

Measurement of ethane as a pure liquid is not very common. Since the liquid has to be handled at a low temperature, it presents unique problems with meters. However, the data for making PVT corrections are available and accepted.

The measurement of ethane liquid mixtures is more common. As noted above, data on pure ethane are readily available, whereas data on ethane-rich streams are limited, and metering accuracy will suffer accordingly. This is particularly true for variable mixtures. At times the better option is to use a densitometer with careful attention to proper sampling to minimize sample errors due to temperature and pressure variations in the stream.

Propane can be handled as a liquid or a gas since its critical temperature is 206°F at a pressure of 616 psia. At normal ambient temperature, it can be a gas or liquid, depending on the pressure. However, to solve measurement problems, the phase relationship must be known so that a single phase of liquid or gas is considered. The meter can then be properly sized, and the two-phase or phase-changing regions can be avoided. As with any fluid, the closer to phase change the measurement is attempted, the more difficult measurement becomes.

Ethylene is a popular hydrocarbon feed stock used in the chemical industry. It is difficult to measure in a number of industrial cases since its critical temperature is 48.5°F at 731 psia, which means that it is handled in an area of greatest density sensitivity (i.e., the measurement problem becomes one of the correct density measurement). Near this point in the vapor phase, the compressibility factor changes as rapidly as 1% per degree Fahrenheit and 0.5% per 1 psi pressure. This problem is significant enough to, at times, require heating prior to flow measurement to obtain sufficient accuracy. Heating to 80° to 90°F will minimize these density changes. This represents an example where the effects of the fluid characteristics on measurement accuracy are so significant that they are changed prior to measurement.

Another characteristic of ethylene is that a very fine carbon dust is produced during processing, which cannot be removed with a 5 micron filter. Depending on the frequency of buildup, the meter and meter piping must be cleaned so that the flow characteristics are not changed. This also affects measurement transducers and their lead lines, which must be cleaned.

Ethylene often contains small quantities of hydrogen. This can affect filled differential transducers; an internal pressure builds up over a period of time to the point that the unit will not be capable of meeting calibration, and replacement is required. How often this happens varies with pressure and hydrogen content.

Ethylene requires that special elastomer materials are used for those parts of a meter that come in contact with the fluid (“wetted” parts). Meter materials should be checked, and meters should be ordered to accommodate these requirements.

Propylene is another popular feed stock for the chemical industry. It is somewhat easier to measure than ethylene. This is true because its critical temperature is 197°F at a critical pressure of 667 psia. Propylene is less reactive with most meter materials, but materials in meter seals should be checked carefully for reaction with propylene.

Carbon dioxide, commonly measured, is used in the oil and gas industry for the recovery of crude oil. It also has a troublesome critical temperature of 88°F at a critical pressure of 1,071 psia that is significant when attempting flow measurement. The compressibility factors in these ranges may represent as much as a 200% correction and will be the controlling factor in achieving accurate flow measurement. Carbon dioxide is quite often handled as a mixture, which further complicates the measurement of density. Data are available from the NIST in Boulder, Colorado, for mixtures of CO₂ from 94 to 99.7% containing small amounts of methane (0 to 2%), ethane (0 to 1%), propane (0 to 2%), and nitrogen (0 to 2%), as well as pure CO₂ (Figure 6-8).

Once again, carbon dioxide measurement is not easy because of the density sensitivity at normal operating conditions—even if well removed from the two-phase flow region. The solution requires continual integration of density with the flow device because of the rapid changes. A computer is needed to do this calculation rapidly enough.

Another measurement problem is CO₂ wetness. If water is present at a sufficient level, a hydrate may be formed at temperatures well above freezing (32°F). In addition, wet CO₂ is very corrosive, and a large amount of corrosion products will move with the gas; this often results in deposits that can cause problems with meters and other equipment such as densitometers. Likewise, CO₂ causes most standard seal materials, such as rubber and Teflon, to break down, and lubricants rapidly deteriorate in the presence of CO₂. All these factors should be taken into account before designing a carbon dioxide metering station.

Steam flow measurement is often one of the most misunderstood of all flow measurements made in industry. There are many reasons for this, but fluid problems are the most important. Measuring steam as a fluid is fundamentally the same as measuring any flowing fluid. If the fluid dynamics and the thermodynamics are known, then the first steps toward accurate measurement have been taken. However, these two areas are not particularly well known or understood in many of the applications in which steam measurement is required.

As commonly used, the term “steam” is meaningless when considering measurement. In the industrial world, the definition narrows somewhat, but even here it covers a variety of flowing conditions. The following sections cover the three possible steam flowing conditions: wet, saturated, and superheated.

Wet (quality) steam is usually the most difficult fluid to measure. A wet steam is a fluid that contains both condensed hot water and steam. In the two-phase portion of the phase diagram (see Figure 6-9), with the same temperature and pressure there is a different density. Therefore a third parameter, quality, must be added to correct a measurement for the right density. Quality is defined as the ratio of percent flow that is steam to the percent flow that is free water, in mass. For example, 95% quality means that 95% of the flow is steam and 5% is water by mass.

Quality can be determined by a calorimeter test (batch operation), which is valid only until changes in the system take place (i.e., flow rate or density change). In addition to these fluid identification problems, the effect of two phases on the meter mechanism can create errors. Because of these problems, quality steam measurement is inaccurate and should be attempted only as a last resort, recognizing that it will have very wide accuracy tolerances.

Saturated steam has no free water present and exists only at one pressure and a corresponding temperature. However, water is not always pure; therefore the exact saturation point, pressure and temperature may be difficult to define. At the same pressure and a higher temperature, the steam is superheated; at a lower temperature, condensation takes place, and the fluid becomes quality steam. Saturated steam exists at a boiler, but when steam leaves the boiler (assuming no superheat is added), the flow creates a pressure drop, and there is the possibility of a temperature drop, depending on flow line insulation. Therefore, steam traveling through a plant will normally be superheated (i.e., the pressure drops, but the temperature is relatively constant and the steam is almost never saturated away from the boiler). Table 6-1 shows the measurements required to determine steam density.

Most designers state that they will be handling “saturated steam” and may not allow for temperature and pressure measurement at the meter. As the flow rate varies, the pressure will change at the meter, and sometimes the temperature will also change. To determine the density, the temperature and pressure must be known (measured) as shown in Table 6-1. If true saturated steam exists at the meter, then measuring the temperature or pressure will define the density.

Wet steam requires that temperature, pressure, and quality are measured to define the density. Saturated steam requires temperature or pressure, and superheated steam requires temperature and pressure. In each case, these measurements must be fed to a computer that calculates the density based on the equations in steam tables.

A problem unique to steam is the large difference between the temperatures of ambient air surrounding a meter and the flowing fluid. This makes the proper measurement of steam temperature a major concern. Without suitable insulation and special precautions, major errors in density will result. In a flowing stream at low velocity, steam tends to stratify by temperature; the steam must be mixed to get the temperature constant across the stream and allow accurate steam measurement (Table 6-1).

In summary, steam is the most difficult fluid to measure accurately. Even with the utmost care, a plant balance of steam and water flows in a power-generating or process plant is very difficult to do well.