Isolating Flow Conditioners Bring Unparalleled Accuracy to Metering Stations

James E. Gallagher, P.E.
Michael P Saunders
Savant Measurement Corporation USA

Accurate gas and liquid measurement is best achieved with an optimized flow profile. All those involved with a custody transfer station benefit from flow profile standards of accuracy that far exceed those of the past 40 years. Well, standards have changed. By including flow conditioning in their latest metering station design standards, the American Petroleum Institute (API) and ISO have recognized a revolutionary new technology that insures an unparalleled degree of flow accuracy. This technology is an isolating flow conditioner-placed upstream of a flowmeter-which conditions the flow such that it enters the flowmeter with a uniform, fully developed profile. This happens regardless of the pipe configuration prior to the conditioner.

Another factor required in this industry is the ability to assess a measurement facility's full cost of ownership. This includes consideration of the initial capital, commissioning, training, spareparts , maintenance, and calibration costs for the equipment' s lifetime. What this means is that full ownership cost is actually several times initial capital investment, spread over time. Considering such costs gives a more realistic financial picture to use as a deciding factor in equipment selection. This, of course, leads also to isolating flow conditioning technology.

Two of the measurement chain's most significant parameters are proper installation and application of flowmeters in conjunction with flow conditioners ; yet, even though they influence the factor s mentioned above, they may be neglected in owners hip cost assessments . This could be a significant over sight, since flow conditioning's role is to ensure that a pipeline' s unpredictably variable flow environment when it enters the custody transfer station is stabilized so it resembles as closely as possible the actual flow of the gas under consideration. The closer this resemblance, the more reliable and fiscally sound the flow measurement.

Installation Effects
All inferential flowmeters (for example, orifice, ultrasonic, and turbine meters) are subject to the effects of velocity profile, swirl, and turbulence structure. The meter calibration factors or empirical discharge coefficients are valid only if geometric and dynamic similarity exists between the metering and calibration conditions or between the metering and empirical database conditions-in other words, under fully developed flow conditions. In fluid mechanics, this is commonly referred to as the Law of Similarity.

In the industrial environment, multiple piping configurations are assembled in series, generating complex problems for standards-writing organizations and flow metering engineers. The challenge is to minimize the difference between the actual flow conditions and the fully developed flow conditions in a pipe, in order to maintain minimum error associated with the selected metering device's performance.

Research programs in both Western Europe and North America have confirmed that many piping configurations and fittings generate disturbances with unknown characteristics. Even a single elbow can generate very different flow conditions-from "ideal" to "fully developed" flow-depending on its radius of curvature (that is, mitered or swept). In addition, the disturbance piping configurations generate is further influenced by the conditions prior to these disturbances.

In general, upstream piping elements may be grouped accordingly:
- Those that distort the mean velocity profile but produce little swirl.
- Those that both distort and generate bulk swirl.

As a result, today's measurement industry focus is to lower uncertainty levels associated with these distorted flow conditions.

Flow Conditioners
The problem, then, is to minimize the difference between real and distorted flow conditions on the selected metering device, thus maintaining the low uncertainty required for fiscal applications . For clarity, this will be referred to as "pseudo- fully developed" flow .

A method to circumvent the influence of the fluid dynamics on the meter 's performance is to install a flow conditioner in combination with straight lengths of pipe to "isolate" the meter from upstream piping disturbances . This isolation, however, is never perfect.

Pseudo-Fully Developed Flow
From a practical standpoint, we generally refer to fully developed flow in terms of swirl-free, axisymmetric, time average, velocity profile in accordance with the Power Law or Law of the Wall prediction.

To bridge the gap between research and industrial applications, the term pseudo-fully developed flow will be defined as follows:
"The slope of the orifice meter's discharge coefficient deviation or meter factor deviation that asymptotically approaches zero as the axial distance from the flowmeter to the upstream flow conditioner increases."

Isolating Flow Conditioner
To truly isolate flowmeters, the optimal flow conditioner, placed in sequence before the flowmeter, should achieve the following design objectives:
- Low permanent pressure loss (low head ratio).
- Low fouling rate.
- Rigorous mechanical design.
- Moderate cost of construction.
- Elimination of swirl [less than 2°-when the swirl angle is less than or equal to two (2) degrees, as conventionally measured using pitot tube devices, swirl is regarded as virtually eliminated].
- Independence of tap sensing location (for orifice meters).
- Pseudo-fully developed flow for both short and long straight lengths of pipe.

For turbine and ultrasonic meters, when the empirical meter factors for both short and long piping lengths are approximately +/- 0.10% for liquid applications, or approximately +/- 0.25% for gas applications, and if it is also shown to be independent of axial position, then it is assumed to be at a minimum and to be pseudo-fully developed.

For orifice meters, the term Cd deviation (%) refers to the percent deviation of the empirical coefficient of discharge or meter calibration factor from fully developed flow to the disturbed test conditions. Desirably, this deviation should be as near to zero as possible. As explained above, a minimal deviation is regarded as +/- 0.25% for gas applications.

Experimental Results
Several flow conditioners have been evaluated by the Gas Research Institute for comparison purposes as part of their Installation Effects Research Program. For these tests, the same test loop or apparatus was used, to provide consistency between experiments.

For the test loop, gas enters a stagnation bottle and flows to a straight section of pipe. The gas then enters a 90° elbow or tee followed by a meter tube and flowmeter. The flow conditioners tested are positioned at various upstream distances, X, from the orifice plate. To obtain dimensionless terms, the distance X was divided by the meter tube nominal diameter, D.

For the experiments, the selected flowmeter was a concentric, flange-tapped, square-edged orifice meter with Betas of 0.67 and 0.75. The internal diameter of the meter tube, IDp, was 102.29 mm (4.027 inches) and the length of the meter tube, L1, was 17 nominal pipe diameters (17D). For certain AGA tube bundle measurements, the length of the meter tube, L1, was increased to 45D and 100D lengths. The flow disturbance was created by either a 90° elbow or a tee installed at the inlet to the meter tube.

Analysis of Results
The results obtained for the AGA design, using meter tube lengths of 17D, 45D, and 100D,
indicate a minimal deviation when:
· L1 = 17D; and X/D = 12 - 15
· L1 = 45D; and X/D = 8 - 9
· L1 = 100D; and X/D = 8 - 9 or > 45

Tests on four flow conditioners in a 17D long test pipe with a tee were funded by GRI. The Beta for the orifice meter was 0.67 and the Reynolds number was approximately 900,000.

These results are not surprising in light of current understanding of pipe flows. The tube bundle-long relied upon to condition the disturbances present in gas flow-does an excellent job of eliminating swirl. However, the fixed diameter tubes generate an unstable turbulence structure that begins to redevelop rapidly. Also, the constant and high radial porosity does not offer a method to redistribute any asymmetric flow patterns.

A new breed of isolating flow conditioners produces pseudo-fully developed flow conditions for both short and long piping configurations. This is evidenced by the slope of the orifice meter's discharge coefficient deviation or meter factor deviation asymptotically approaching zero as the axial distance from the flowmeter to the upstream flow conditioner increases. The new breed of flow conditioners has also demonstrated an insensitivity to tap sensing location, confirming the presence of pseudo-fully developed flow.

Measurement Standards
Orifice Meters
The research programs have clearly indicated that the requirements specified in both orifice standards are erroneous and that minimum straight length specifications in the standards (ISO 5167 and AGA no. 3) are in urgent need of revision.

Present domestic and international measurement standards provide installation specifications for pipe length requirements and flow conditioners upstream of orifice meters (ANSI's 2530 and ISO's 5167). A significant revision with respect to piping configurations with and without flow conditioners is presently underway for both standards. Both standards are out for ballot in 1999.

With respect to installation effects and the near-term flow field, the correlating parameters that impact similarity vary with flowmeter type and design. However, it is generally accepted that the concentric, square-edged, flange-tapped orifice meter exhibits a high sensitivity to time average velocity profile, turbulence structure, and bulk swirl and tap location.

In North America, current design practices utilize short upstream piping lengths with a specific flow conditioner-AGA tube bundles-to provide pseudo-fully developed flow in accordance with the applicable measurement standard (ANSI 2530/A.G.A. 3/API MPMS 14.3). Most North American installations consist of 90° elbows or complex header configurations upstream of the orifice meter. Tube bundles in combination with piping lengths of 17 diameters (17D) have been installed to eliminate swirl and distorted velocity profiles. Ten diameters (10D) of straight pipe are required between the upstream piping fitting and the exit of the tube bundle, and 7 diameters (7D) of straight pipe are required between the exit of the tube bundle and the orifice meter.

Recent research indicates that the flow conditioning error is a function of time-averaged velocity profile, swirl angle, tap sensing location, and turbulence structure. As a result of these new findings, a significant improvement in flow conditioner performance has been achieved over other devices designed to tackle velocity and swirl alone.

Ultrasonic Meters
Ultrasonic meter technology is relatively new to fiscal applications. This technology shows tremendous potential for performance equal to or better than most world-class flow calibration laboratories.

Preliminary research in natural gas has indicated the need for flow conditioners to ensure compliance with the Law of Similarity and an uncertainty of +/- 0.25%.

Preliminary research in liquid has also indicated the need for flow conditioners that ensure compliance with the Law of Similarity and an uncertainty of +/- 0.1%.

Turbine Meters
For gas applications, AGA report no. 7 and ISO 9951 cover the requirements for their installation and performance.

For liquid applications, API MPMS chapters 5 and 6 cover the requirements for their installation and performance. Recent research in the laboratory and the field has indicated the sensitivity to velocity profiles approaching the turbine meter. Errors of as much as 1.0% were reported due to partially blocked strainers and/or provers located upstream of the meter runs.

Outlook
Designing and operating an accurate flowmeter application requires understanding of the fluid's physical properties. An envelope must be drawn around the process (or operating) conditions, and the identification of any special conditions. Understanding the physical principles upon which the selected flowmeter is based and comprehending its sensitivities to physical and process conditions is critical. Most important, designing and operating an accurate measurement facility requires compliance with the Law of Similarity, which is what the flow conditioner insures. Placing an isolated flow conditioner prior to the flowmeter will condition the disturbed fluid entering the conditioner so that it proceeds to the flowmeter with a virtually ideal bullet- shaped profile for extremely accurate measurement. Managers who examine these factors will discover their estimate of full cost of ownership is not only more accurate but will positively