Flow Control: Technology information & new products for fluid handling engineers

February 2010

Back to List | Print Article

Gas Flow Check Metering
Benefits of Pipe Waveguide Method for Clamp-On Ultrasonic Flowmeters

by Ron McCarthy

Gas and oil are different in their physical, chemical, and ultrasonic properties. Nevertheless, each fluid state can benefit from new technology that uses a pipe wall as a sonic waveguide for accurate, repeatable, and reliable non-intrusive flow measurement. The following article explains how clamp-on ultrasonic flowmeters featuring such technology can be used to provide highly accurate check metering.

All transit-time ultrasonic meters rely on the simple principle that sound moves at a lower velocity when traveling against the direction of a fluid and at a higher velocity when traveling with the fluid. By precisely measuring the times of the sound transmission in each direction (upstream and downstream), the line integrated fluid flow velocity can be computed. For an insert or wetted ultrasonic meter, where the transducers are inserted through a port into the spool, the equation for flow could be written as shown in equation (1) with reference to Figure 1.

Figure 1. Model for transit-time equation

Although this simple relationship has been known for more than a century, it was not until the introduction of accurate, stable time measurement technology and advanced ultrasonic transducers that this theory could be practically applied to gas flow measurement. Piezoelectric transducers are the most common type of transducers used for transit-time ultrasonic flowmeters. The piezoelectric ceramic is used to convert electronic signals into sound energy and inversely convert sound energy back into an electronic signal. This material must be stable to ensure that minimal “mismatch” exists between the two elements used in the transducer pair, otherwise this mismatch would be perceived by the flowmeter as a small flow offset.

The wetted transducer configuration is able to inject an ultrasonic beam across various chords along the meter body’s cross-section. Depending on the number of acoustic beams, this configuration can improve flow profile averaging. This can cause problems if the transducer ports are blocked by contaminates in the gas pipeline, thus impeding the acoustic beam and potentially interrupting operation.

Non-Intrusive, Clamp-On Flowmetering
For flow survey, check metering, or installations where cutting into an existing line is not practical, clamp-on flowmeters have become the meter of choice. Due to recent advances in transducer design and signal processing technology, clamp-on meters using pipe waveguide technology are now demonstrating accuracy and repeatability that rivals inline transducer meters. The natural ultrasonic mode of this new technology is represented by the following equation:

As the sonic wave is introduced to the pipe wall, it travels down it axially between the transmit and receive transducers, maintaining a coherent signal wave shape as illustrated in Figure 2.

Figure 2. Sonic wave signal maintains coherence

Shear mode clamp-on transducers do not match the waveguide properties of the pipe, and will quickly lose signal coherency after traveling only a short distance along the pipe wall, as shown in Figure 3. This signal does not provide the coherency required for operation over a wide range of fluid properties, or for high-velocity gas applications.

Figure 3. Shear mode signal loses coherence

Pipe Waveguide Clamp-On Technology Improves Accuracy
Pipe waveguide clamp-on operation is less affected by beam blowing conditions found in gas applications due to transducer size and the use of the pipe as a waveguide.

In clamp-on ultrasonic technology, Snell’s Law of Refraction determines the place where the beam actually emerges from the pipe wall as it travels through the gas, and is independent of the actual spacing of the transducers. The angle of emergence is dependent on the sonic propagation velocity of the fluid relative to the propagation velocity of the beam in the pipe wall, as described in equations (3) and (4) and Figure 4.

Figure 4. Refraction of clamp-on beam.

High flow velocities can cause significant displacement (or blowing) of the sonic beam as it travels from the transmit to receive transducer. This is primarily a concern for gas metering applications where the gas flow velocity represents a relatively high percentage of the gas sound speed (VOS) 1350ft/sec. In insert (pencil beam) transducer ultrasonic meters, the beam can actually be blown away from the receive transducer. In pipe waveguide ultrasonic flowmeters, it simply changes the place on the pipe wall from which the beam emerges, permitting operation at high flowrates. This beam displacement, where the upstream and downstream receive signals originate from different locations along the pipe wall, is illustrated in Figure 5.

Figure 5. The effect of high velocity flow on upstream and downstream sonic beam position.

In shear mode clamp-on ultrasonic meters, as the flow increases, the upstream and downstream receive signals emanate from different positions on the pipe wall relative to the transmitting transducer. This makes the sonic signals so dissimilar that they lose any correlation features. Without proper signal correlation, it is impossible to accurately measure the transit-time and provide stable operation or accurate flow measurement.

The pipe waveguide-based signal transmission method, on the other hand, makes the sonic signal independent from where on the pipe wall it has emerged, and assures the phase coherency of the receive signal. This permits both stable operation and accurate measurement at low and high flow velocities, as well as the ability to operate in varying pressure applications.

Clamp-On Transit-Time Flow Equation
Clamp-on transit-time systems differ from insert systems in the way sonic energy enters the fluid medium. Insert transducers inject their acoustic signal perpendicular to their emitting surfaces, whereas clamp-on systems must inject their acoustic signals at an angle to the emitting surfaces to allow for sonic beam refraction. The pipe wall injects the acoustic signals directly into the fluid, which travels in a vector parallel to the pipe wall.

The clamp-on acoustic beam must traverse the pipe across its diameter as the beam must be at right angles to the pipe. The vector analysis for the clamp-on geometry results in the simple equation shown below, which uses the transducer phase velocity Vφ shown earlier in Equation 4.

The transit-time flowmeter measures Δt, by subtracting T_UP and T_DN. It must also derive the required time-in-the-fluid “T_Fluid” (T_F) and sound velocity “V_S”. Sufficient information is available to compute (Theta), the system calibration angle, and the time of the sonic beam traveling down the pipe wall to accomplish this in real time. This can be achieved by solving the acoustic timing equation (7) below, or by generating an appropriate lookup table. Finally, the line integrated flow velocity is then compensated for flow profile effects, which requires only inputting the installation pipe configuration (i.e., elbow, reducer, etc., and the number of diameters downstream from such a configuration). The profile averaged flow velocity is converted to volumetric flowrate using the correct pipe inside diameter.

Using The Pipe Wall Signal For Zero Drift Correction
A pipe waveguide-based transit-time meter is capable of measuring a “live” or active zero flow reference. All transit-time meters are affected to some degree by zero drift or zero flow offset. The use of this feature eliminates both the need to actually “set” a zero flow adjustment in the field or suffer from zero drift in this type of flowmeter. This zero drift can be attributed to subtle changes in the flowmeter electronics and/or changes in the temperature of the ultrasonic transducers. This is why we temperature compensate the transducers. Since the difference in transit time is very small, this is a potential problem area in other ultrasonic flowmeters.

The most common cause of zero drift is non-common mode temperature that can occur between the transducer pair. Sun loading is one situation that can cause differential heating of the transducer pair. Some users install their ultrasonic flowmeters under the protection of a shed to avoid such sun loading.

Figure 6a. Acoustic signal splitting

The pipe wall can also be viewed as an acoustic beam splitter, where one portion of the acoustic energy travels into the fluid and the other straight down the pipe wall into the receive transducer (Figure 6a). A typical receive signal for this configuration is shown below in the Figure 6b oscilloscope trace meter installation using pipe waveguide transducers.

The beam that travels through the fluid is used to measure flowrate directly, but the beam traveling only along the pipe wall never “sees” the liquid and is therefore unaffected by flowrate. The pipe waveguide transmission technique takes advantage of this pipe wall signal, which is digitally processed in the same manner as the fluid signal. By continually measuring the transit-time difference of the pipe wall signal, the meter dynamically corrects the flow measurement for any changes in zero drift. This method is valid whether the drift is due to changes in the electronics, cables, or transducers. Not only does this improve accuracy during normal operation, but it also permits accurate low-flow operation and can reduce the need for multiple meters to cover a wide range of flowrate.

This enables automatic correction of the zero flow set point, even if flow cannot be stopped. Non-common mode environmental conditions vary hour by hour. The function for automatic zero correction only works if both transducers are installed on the same side of the pipe in a reflect mode. It cannot be implemented for either chordal (insert) or shear mode clamp-on systems. In chordal systems, no zero pipe wall reference path exists; and in the shear mode system, the upstream and downstream pipe wall signals cannot be correlated.

Benefits of Diametric Reflect Clamp-On Transducer Installation
As with most conventional meters, the flow profile imposed on ultrasonic meters is dependent on upstream piping configuration. Depending on the number of ultrasonic beams used, an ultrasonic meter is affected to a smaller or larger degree by upstream conditions.

Figure 6b. Typical pipe waveguide signal

An ultrasonic flowmeter installation, in which the transducers are installed on opposite sides of the pipe (Direct Mode), can be seriously affected by cross-flow. The susceptibility of this geometry to cross-flow profile (such as that produced after a single upstream elbow) can generate flow errors in the range of 3 percent to 6 percent per degree for gas applications. Figure 7 depicts the mechanism for this error and how an installation on the same side of the pipe (reflect) can compensate for cross-flow condition.

Figure 7. Cross-flow diagram

With transducers installed on opposite sides of the pipe (acoustic path AB), the beam geometry cannot compensate for the error caused by the cross-flow angle. However, when the transducers are installed on the same side of the pipe (acoustic path ABC), the geometry provides compensation (0.007 percent per degree of cross-flow). In this case, path AB would tend to under-read and path BC would over-read, thereby providing a self-compensating and correct average flow indication.

Insert and clamp-on installations, with transducers installed in the AB configuration, are more susceptible to cross-flow conditions. (As described by Snell’s Law). Accordingly, pipe waveguide systems use normally diametric reflect installation.

Other profile aberrations, such as swirl induced by single and double out of plane elbows, can also impact meter accuracy. In the case of insert transducer flowmeters, swirl tends to align directly with the chordal paths and can produce errors in the individual chord. Compensation for this error can be accomplished by using multiple chords.

Pipe waveguide-based flowmeters are not as susceptible to swirl because the swirl is at right angles to the diametric beam, and therefore does not induce a time difference. Swirl can also impact the accuracy of a pipe waveguide-based system, since it flattens the shape of the flow profile (Figure 8). Errors up to 5 percent could be induced from the flow profile flattening effect of Swirl. With a better understanding of this phenomenon, the proper compensation can be input to correct for the majority of these errors.

Figure 8. Flow profile flattening due to a 90-degree elbow.

The pipe waveguide-based clamp-on transit-time equations and functions described earlier apply equally to the measurement of any liquid or gas. Gas metering does, however, present manufacturers with a far greater challenge than most liquid metering applications.

The sonic impedance of liquid, which determines how much sonic energy in the pipe wall is transferred into the flow stream, is much higher than for gas applications. This is because the sonic impedance is the product of its density and its sonic propagation velocity. Liquid is much higher in both categories, and it is much easier to transfer sonic energy into a liquid than a gas. As the pressure of a gas increases, so does its density; although it becomes more accepting of sonic energy, it never reaches the magnitude of any liquid signal level. Table 1 illustrates the difference in acoustic impedance between water and natural gas at various operating pressures. Also shown is the signal amplitude and beam angle.

Table 1. Relative acoustic impedance of water and gas.

It is important to note that the detection of upstream versus downstream transit times is independent of the signal amplitude, the effects of low-pressure causes low signal amplitude and reduces the signal-to-noise ratio. This increases data scatter, which is integrated out in normal applications. At normal transmission line pressures of between 400 PSI to 1,200 PSI, the signal is strong and in the millivolt range. In normal application conditions, signal-to-noise ratios of over 100-to-1 are obtained.

Effect of Residual Pipe Noise on Clamp-On Natural Gas Meters
New, improved signal amplifiers can operate on acoustic receive signals smaller than 0.1 millivolts. However, if the signal baseline contains high levels of synchronous noise, the gas signals can become indistinguishable from the baseline. The elimination of synchronous pipe noise is one of the greatest challenges of clamp-on gas meters.

The low acoustic impedance of gas means that most of the energy will remain in the pipe wall, with only a small fraction entering the gas medium. This sonic pipe wall energy travels along the axis of the pipe, and can continue to reflect back to the receive transducer from flanges and welds until it slowly decays. To prevent pipe noise from obscuring the gas signal acoustic, damping must be used to absorb the pipe wall sonic energy.

A clamp-on system installed on a short pipe section or spool provides a good example of how synchronous pipe noise can interfere with the signal. Figure 9a, above, shows the receive signal from a six-inch class 600 clamp-on gas spool meter pressurized to 100 PSIG with transducers installed on the bare pipe surface. The larger saturated signal to the left side of the trace is the main pipe noise signal that continues to decay as time progresses. The actual gas signal is so small that on this scale it is buried in this noise. Figure 9b represents the exact same installation, except the outside surface of the spool is coated with a sound-absorbing material. Even at low pressures, the gas signal now becomes distinct with a high signal-to-noise ratio. Installation of this sound-absorbing material is simple and can be performed on a field installation in less than 15 minutes.

Figure 9a. Synchronous pipe noise on six-inch gas spool hides gas signal.

Figure 9b. Pipe noise elimination permits the gas signal to be seen.

Gas Check Metering
Field-installable, nonintrusive gas flowmeters provide sufficient accuracy to perform check metering, thus offering a great value to those who buy and sell natural gas. The availability of calibrated pipe waveguide clamp-on spools for high accuracy check metering is another option. This design and its performance have been confirmed by testing at major laboratories worldwide. These spool meters also offer low maintenance.

Conformance to Ultrasonic Flowmeter Standards
While clamp-on technology is different from insert technology, it conforms to the same standards, which have been developed worldwide for the ultrasonic meter. They are able to measure the sonic velocity (VOS) to within AGA-10 standards. The spool meters are in conformance with AGA 9.

Clamp-On Flowmeter Transducer Assemblies
Field installation of clamp-on transducers can be performed quickly for temporary
survey jobs or temporary check metering. However, if permanent field installation is required, a more rugged and secure transducer housing is recommended. These housings can be either tack welded to the pipe, or strapped to the pipe as preferred by the user (Figure 10). Transducer spacing is precise and controlled by use of an indexed spacer bar, which prevents any movement between transducers. The clamping assembly applies all the force necessary to prevent any movement of the transducers. The entire assembly is enclosed to prevent accidental disturbance.

Figure 10. Four beam transducer installation

Test Data. Results and setup of the test of 10-inch Siemens clamp-on gas spool flowmeter.

Field & Laboratory Tests
The previous page presents test data for clamp-on ultrasonic meters for check metering purposes. Calibrations show performance from 0.2 percent to 0.45 percent over a large turndown ratio.

In check metering applications, tests for field-installed clamp-on flowmeters in proximity to calibrated reference meters show conformance from 0.25 percent to 1.0 percent for acceptable upstream conditions. In the tests, no attempt was made to correct the clamp-on meters to account for the influences of the pipe configuration flow profile effects. It was always assumed that the reference meter was perfect in its calibration.

For clamp-on gas meter testing, the measured VOS played a key role in validating the installation before any flow comparison was made. In some cases, pipes were found to be out of round, or the assumed inside diameter or wall thickness was found to be incorrect.

The theoretical and practical information presented here suggests the following conclusions are valid:

The field-installed pipe waveguide-based clamp-on ultrasonic flowmeter can be used for check metering with only the knowledge of the Pipe Diameter, Pipe Material and Wall Thickness.
Pipe configuration tables permit substantial increase in accuracy for imperfect installations.
Automatic zero correction functions permit solid accuracy for applications where low flowrates are prevalent. This can help reduce the need for multiple meters to cover wide flow ranges.
The pipe waveguide principle is resistant to beam-blowing effects, which tend to limit the maximum flowrate of conventional insert and shear mode clamp-on meters.

Ron McCarthy is the business development manager for the gas market at Siemens Energy & Automation. Mr. McCarthy currently serves as a member of various industry organizations and working groups, including the Institute of Petroleum (UK), Society of Petroleum Engineers, API ISO Committee, PRCI, AGA Committee, and the GRI Technical Advisory Committee. He has 30-plus years of experience in engineering and sales in both liquid and gas measurement. Mr. McCarthy can be reached at 281 240-4360 or ronald.mccarthy@siemens.com.

Acknowledgments
Siemens acknowledges the work of the calibration labs TransCanada Calibration, PTB, and Southwest Research Institute; whose excellent work contributed to the test results gathered for this article. Also, to the research conducted by working groups within the Pipeline Research Council International.

Back to List | Print Article

Get a FREE Subscription to Flow Control magazine | Subscribe to the Flow Control E-Newsletter

Site Developed & Maintained by the Pinnex Group, LLC

Privacy Policy