Matching the Flowmeter to the Application

Figure 1. The orifice plate is a popular DP element.

Two significant specifications for a flowmeter are accuracy and turndown (or rangeability). The accuracy of all flowmeters depends to some degree on the circumstances of their application. These effects include pressure, temperature, fluid dynamic influences, and external influences. Pressure and temperature cause dimensional and physical changes. Fluid flow and pipe roughness can shift meter performance. Mass meters and displacement meters are the least affected, while turbine, vortex, and ultrasonic meters are the most affected. The upshot is that a traceable and complete calibration in a lab is no guarantee that the same performance and accuracies will be obtained or maintained in the field.

A flowmeter’s turndown is the ratio of the maximum to the minimum flowrate capable of being measured at a specified accuracy, providing a measure of rangeability. High values of turndown can be irrelevant because typical processing flow velocities often fall within relatively narrow ranges. For example, most practical liquid pipeline flowrates in the processing industries range from 0.5 to 12 ft/sec, a turndown ratio of 24-to-1. Lower rates are difficult to measure accurately and higher rates result in high pressure drops, pumping energy costs, and erosion (if solids are present). In the case of pipelines carrying gases in the processing industries, the practical flow velocities range from 15 to 200 ft/sec — a turndown ratio of about 13-to-1. Many actual applications have flowrate ranges well within these extremes.

Figure 2. The Venturi DP element minimizes pressure loss.

Instrument engineers commonly apply volumetric flowmeters to measure liquids. When the process fluid is a gas or steam, measuring mass flow becomes especially important. Because these fluids are compressible, their density changes with pressure and temperature. Volumetric flow measurement is often meaningless in these cases. But volumetric flowmeters work if the application also includes measurement of temperatures and pressures at the meter or these values are known quantities. Simple calculations based on the three variables can infer mass flowrates.

Differential-Pressure Flowmeters
Flowmeters based on differential pressure insert an element within the piped flow. The unrecoverable pressure loss across the element — typically an orifice plate, Venturi, nozzle, pitot tube, or wedge — is a measure of the volumetric flowrate.

Figure 3. The nozzle has Venturi-like properties.

Flowmeters based on differential pressure represent a popular choice in the processing industries, constituting nearly 30 percent of the installations. They have good application flexibility since they can measure liquid, gas, and steam flows and are suitable for extreme temperatures and pressures with moderate pressure losses. Accuracy ranges from 1 percent to 5 percent. Compensation techniques can improve accuracy to 0.5 percent to 1.5 percent.

On the other hand, DP flowmeters are expensive to install and have limited rangeability (turndown) compared to other types. They require a separate transmitter to compute the volumetric flowrate and develop a standard signal. Changes in density, pressure, and viscosity can significantly affect accuracy of DP flowmeters. And while they have no moving parts, maintenance can be intensive.

Orifice plate: Instrument engineers can employ a variety of elements to create the differential pressure, with the orifice plate being the most common. Orifice plates are inexpensive and available in a variety of materials. The turndown ratio, however, is less than 5-to-1, and accuracy is moderate — 2 to 4 percent of full scale. Maintenance of good accuracy requires a sharp edge to the upstream side. This edge will wear and degrade over time. Pressure loss for orifice plates is high relative to other DP elements.

Figure 4. The pitot tube converts fluid kinetic energy into pressure.

Venturi: This element finds use primarily in water and wastewater applications. It handles dirty fluids and does not require upstream flow profiling. Pressure loss is minimal, making it a good choice when little pressure head is available. Rangeability, while better than orifice plates, is less than 6-to-1, with accuracy of +/-1-to-2 percent of full scale. Viscosity effects on accuracy are high. Flow must be turbulent (Reynolds numbers > 10,000). The Venturi has limited acceptance within the processing industries.

Nozzle: Characteristics of nozzle elements mimic those of the Venturi. Nozzles come in three types: ISA 1932 nozzle (common outside of the United States); the long radius nozzle; and the Venturi nozzle.

Figure 5. The wedge DP element handles tough, dirty fluids.

Pitot tube: This element is another low-cost DP element used to measure fluid flow, especially air flow in ventilation and HVAC systems. Pitots commonly measure the speed of airplanes. This element works by converting the kinetic energy of the flow velocity into potential energy (pressure). Engineers can easily insert the pitot tube into existing piping, minimizing installation costs. It cannot handle dirty flow and has limited rangeability.

Like an insertion meter, the pitot tube makes a measurement at a point within the pipeline or ductwork. An exception is the annubar pitot tube, which contains multiple orifices. This element sees the dynamic pressure across the velocity profile, providing an averaging effect.

Figure 6. Electromagnetic flowmeters require conductive fluids.

Wedge: This element consists of a V-shaped restriction molded into the meter body. In profile the wedge looks like a segmental orifice plate to the incoming fluid, but it’s more expensive than an orifice plate. This basic meter has been on the market for over 40 years, demonstrating its ability to handle tough, dirty fluids. The slanted faces of the wedge provide self-scouring action and minimize damage from impact with secondary phases. Rangeability of 8-to-1 is relatively high for a DP element and accuracies are possible to +/-0.5 percent of full scale.

Figure 7. AC magmeters have good signal-to-noise ratios, but require frequent calibrations to eliminate zero drift.

The wedge flowmeter, however, is not approved for flow measurement by the American Gas Association (www.aga.org) or the American Petroleum Institute (www.api.org). Despite this, it’s a popular choice in oil and gas applications, especially in production fields. For difficult fluids it can be equipped with a pair of remote seals that effectively isolate the metered fluid from the DP transmitter without affecting accuracy.

Electromagnetic Flowmeters
Electromagnetic flowmeters (magmeters) are a popular choice among instrument engineers, making up about 20 percent of flowmeter installations. Faraday’s law says that a conductor moving through a magnetic field produces an electric signal. In this case the fluid is the conductor and electromagnetic coils surrounding the meter body generate the magnetic field.

Figure 9. Coriolis mass flowmeters are ideal, but relatively costly.

Magmeters have no moving parts. They offer wide rangeability and an unobstructed flow path. Magmeters are ideal for slurries, and perform well with corrosive and erosive fluids. Engineers can pick a nonconductive liner material for the flow tube — Teflon, Tefzel, rubbers, ceramic, etc. — to suit the fluid measured. Magmeters require minimum lengths of straight pipe to condition the flow profile. On the downside, the fluid must be conductive, which rules out most petroleum-based flows. Magmeters also have physical pressure and temperature limits.

Figure 10. Thermal mass flowmeters excel at measuring low flow rates.

Magmeters have rangeability up to 1000-to-1, depending on the maximum tolerable measurement error. For flow velocities of 0.5 to 50 ft/sec, accuracies are usually stated as a percent of rate, for lower velocities, accuracies are stated as a percent of span. Typical velocities measured range from three to 15 ft/sec for water and clean chemicals, three to six ft/sec for abrasive fluids, and six to 12 ft/sec for coatings and liquids with entrained air.

The voltage developed across the electrodes is in the millivolt range and should be carried via shielded cable to a nearby converter. The converter near the flow tube boosts the signal to a standard (four to 20 mA) or frequency output (zero to 10,000 Hz) for transmission to a display or controller.

Figure 11. Vortex meters have no moving parts, but may be affected by vibration.

The coils producing the magnetic field may be excited by AC, DC, or pulsed DC. DC excitation accounts for most magmeter installations today. With AC excitation, line voltage powers the magnetic coils, and the resulting signal has a sine wave shape of line frequency whose amplitude is a linear function of flow velocity. AC designs produce a high coil current (typically 3.2 amp) and a good signal-to-noise ratio. They are relatively insensitive to media noise. However, they consume much power and can be plagued with zero drift caused by eddy current noise. To compensate, operators stop the flow and set the transmitter output to zero. Frequent operations to eliminate eddy current noise are sometimes necessary to maintain accuracy, which is about +/-1 percent of rate.

Figure 12. Turbine meters are well-accepted in the processing industries.

With DC excitation designs, a lower frequency DC pulse excites the magnetic coils. The converter reads both the flow and noise signals during a pulse. In between pulses, however, it sees only noise, permitting noise cancellation after each cycle. So zero drift is not a problem for pulsed-DC designs. These magmeters offer 0.15 percent to 0.5 percent accuracy and low power consumption. The low coil current (0.1 to 0.5 amp) comes with smaller signal and low signal-to-noise ratio. Low frequency designs have longer response times.

Figure 13. Doppler ultrasonic flowmeters work best on slurries and dirty liquids.

A newer DC design develops a stronger flow signal by increasing the DC coil excitation current (0.7 to five amp). This design improves the signal-to-noise ratio, but increases power consumption and cost. Flow tubes must generally be less than 20 inches in diameter. Another new design employs dual excitation, pulsing the coils at about seven Hz for zero stability and also at 70 Hz for a stronger signal and improved response times (0.1 sec). But this design raises costs and applies only to flow tubes smaller than 16 inches in diameter. Eddy current instability increases proportionally as sensor size increases.

Mass Flowmeters

Figure 15. Transit-time ultrasonic flowmeters work best on clean fluids.

Mass flowmeters directly measure mass flowrate, as compared to other flowmeter types, which measure flowrate based on volumetric principles. The two primary types of mass flowmeters, accounting for about 18 percent of installations, are Coriolis and thermal mass flowmeters.

Coriolis flowmeters represent the ideal flowmeter. They can be applied to virtually any fluid — liquids, gasses, and slurries. They are immune to changes in viscosity, pressure, temperature, or density of the piped fluid and are unaffected by flow profile disturbances. They offer high accuracy (+/-0.1 percent) and rangeability (100-to-1), and provide an additional measurement of fluid density. But their purchase and installation costs are relatively high, as well as their sensitivity to vibration. In addition, they come in a limited selection of sizes, rarely over two inches.

Figure 16. Positive-displacement flowmeters are unaffected by poor flow profiles.

In these meters, the fluid measured runs through one or two vibrating tubes. Tube designs take on a multitude of shapes, depending on the manufacturer. Coriolis forces resulting from the vibration and flowing fluid act on the tubes, causing them to twist or deform. Sensors pick up the change in tube shapes. Sensor signals may be related directly to mass flowrate.

The thickness of the tubing walls varies considerably from design to design. But even the sturdiest tubing will be thinner than the process piping. Long, bent tubes twist more easily than do short, straight tubes, and so they will generate stronger signals under the same conditions. Curved-tube designs provide wider rangeability (up to 200-to-1), while straight-tube meters are limited to 30-to-1 to 50-to-1, with lower accuracy. Straight-tube meters are more immune to pipeline stresses and vibration.

Thermal mass flowmeters add heat to the flow stream to make measurements. One or more temperature sensors measure how quickly this heat dissipates. One technique maintains a heated sensor at a constant temperature, measuring the current required. Another way is to measure the temperature difference between a sensor in the flow stream and a heated sensor outside the stream. Higher flowrates result in increased cooling. The cooling effects are a function of mass flow.

These meters are compact, and offer wide rangeability (50-to-1) and good repeatability. They have a fast response time, especially for gasses, and excel at measuring low mass flowrates. They assume a constant specific heat and thermal conductivity of the fluid. Changes in density cause calibration shift, and coating of the sensor can cause drift. Additionally, they measure a point within the pipe diameter, so the flow profile must be known. Accuracy levels typically range from 1 percent to 3 percent. However, manufacturers expect to improve their accuracy, which will widen their application.

Vortex/Swirl Flowmeters
Place a bluff body within a fluid pipeline and the stream will alternately produce whirling vortices on either side of the body. The spacing between the vortices remains constant, regardless of the fluid velocity. Within a pipeline, the vortice effect attenuates a few pipe diameters downstream from the bluff body. Typically, piezoelectric or capacitance sensors count the pressure oscillations caused by the vortices. The vortex-shedding frequency provides a measure of fluid velocity.

Like a vortex meter, swirl meters detect pressure pulsations, whose frequency depends on the flowrate. Vanes within the meter cause the incoming fluid to take the shape of a rotating helical coil of relatively low-pressure zones. At low flowrates the low pressure swirls or coils are far apart. As flowrates increase, the coils become compressed, coming closer together. Detectors within the meter measure the frequency of the swirling low-pressure zones as they pass through the meter. A deswirler at the meter’s exit eliminates any effects on downstream instruments.

Engineers can apply these meters to liquid, gas, or steam, but should avoid dirty, corrosive, or abrasive fluids. The meters have no moving parts, have wide rangeability, and are immune to pressure, temperature, or density changes. These meters, however, lose effectiveness as viscosity increases. They are also affected by vibration or pulsation.

Swirl meters are less sensitive to flow profiles than vortex meters. While vortex meters may require a minimum straight-pipe run of 15 or more pipe diameters upstream and five pipe diameters downstream, swirl meters require only three diameters upstream and one diameter downstream. Both may be installed horizontally or vertically, but must be kept full of fluid during operation. About half of all vortex meter installations require “necking down” oversized process piping by concentric reducers and expanders.

Turbine Flowmeters
A piped fluid that acts on turbine vanes will rotate the turbine with a rotational rate proportional to flowrate. Turbine meters make up about 8 percent of the installed base. Rangeability can reach 100-to-1 if the meter measures the rate of a single fluid at constant conditions. Accuracy is high and may top +/-0.1 percent.

Turbine meters are suitable for extreme temperatures and pressures and can be applied to liquids and gasses. They install easily and are well accepted in processing applications, especially the oil and gas industries. On the other hand, turbine meters suit only low viscosities and are sensitive to flow profile and vibration. Having moving parts, they usually require frequent calibration. Partially open valves upstream from a turbine meter can cause significant errors. Typically, turbine meters require flow upstream straightening vanes.

Additional considerations include:
• Possible flashing or cavitations from pressure drops. Flashing causes the meter to read high while cavitations result in rotor damage. The downstream pressure must equal 1.25 times the vapor pressure plus 2X the pressure drop.
• Small amounts of air entrainment (<100 mg/l) will make the meter read high, while large quantities can destroy the rotor.
• Solids entrained in the fluid can damage the meter. In these cases, a flushing strainer or filter must be installed upstream of the flowmeter.

Ultrasonic Flowmeters
Doppler flowmeters, limited to liquids, send an ultrasonic beam into the fluid and look at reflections from discontinuities such as entrained particles or bubbles. Reflections from the moving objects will experience a frequency shift that is a measure of the flowrate. Transit-time flowmeters have beam transmitters and receivers separated by a known distance. They look at the differences in time of flight for ultrasonic beams moving with and against the flow. Doppler meters work best on slurries or dirty liquids, while transit-time meters work best on clean fluids.

These flowmeters have no moving parts and offer an unobstructed passage for fluid flow. Ultrasonic flowmeters have wide rangeability, and some designs can simply be strapped on to the outside of pipeline. Typical accuracy of these clamp-on flowmeters is about 5 percent. Clamp-on units generally will not work well if the pipe has a lining or if the pipe material is stainless steel, PVC, or fiberglass. In addition, the inside of the pipe must be free of sound-absorbing material, such as dirty grease or scale. Multi-beam, transit-time units in direct contact with the fluid, however, now have sufficient accuracy to be used for custody transfer of gases and liquids.

Published data on ultrasonic flowmeters is limited. They have a relatively high initial cost and performance is sensitive to multiple installation factors. Deposit buildup on the pipe walls can significantly affect accuracy, as can extreme vibrations.

Variable-Area Flowmeters
Commonly known as rotameters, these devices provide practical solutions for many flowmetering applications. They are simple and inexpensive, performing as well or better than many high-tech flowmeters on the market. Every rotameter basically consists of two components — a tapered metering tube made of glass, metal, or plastic and a float that rides within the tube. These components come in a wide variety of shapes, sizes, weights, and materials of construction, making them adaptable to a wide range of application needs. They are easy to install and maintain, but must be mounted vertically and plumb. Some designs can handle high pressures and viscosities. Today, most manufacturers have taken all the factors and data involved in sizing a rotameter and incorporated them in a sophisticated software program.

Accuracy, however, is relatively low (+/-2 percent) and depends on precise knowledge of the fluid and process. Rotameters are susceptible to vibration and plugging by solids and are affected by density and temperature changes. They apply primarily to flowrates below 200 GPM and pipe sizes less than three inches.

Target Flowmeters
This meter inserts a physical target within the fluid flow. The moving fluid deflects a force bar attached to the target. The deflection caused by the force depends on the target area, as well as the fluid density and velocity. These flowmeters have no moving parts such as bearings and can measure fluid flowrates for any line size above 0.5 inches. Engineers can adapt the flowmeter to different fluids or flowrate ranges by changing the target size and material. Rangeability is about 15-to-1. One disadvantage is that the target meter’s calibration must be verified in the field.

Positive-Displacement Flowmeters
Precision-fitted rotors measure fluid flow in positive-displacement flowmeters. The volumes between the rotors are known and fixed. The rotor rotation varies directly with the volume of the fluid being displaced. Counters monitor the number of rotations and a transmitter converts the counts to volumetric flowrate. Principal types include:
• Single or multiple reciprocating piston meters.
• Oval-gear meters with synchronized, close-fitting teeth.
• Movable nutating discs mounted on a concentric sphere located in spherical side-walled chambers.
• Rotary vanes creating two or more compartments inside the meter’s housing.

These flowmeters are applicable to a wide range of nonabrasive fluids, including those with high viscosities. Examples include heating oils, lubrication oils, polymer additives, ink, and many more. They are unaffected by poor flow profiles. Accuracy may be up to +/- 0.1 percent of full scale with a rangeability of 70-to-1 or better. Moreover, they require no power and can handle high pressures.

Positive-displacement flowmeters cannot handle solids, entrapped air in liquids, or entrained liquids in gases. They are expensive to install and maintain, having many moving parts. Pressure drop across the meters is high.

This is the fifth article in a five-part series on the history and operation of flowmeter technology.

Greg Livelli is a senior product manager for ABB Instrumentation, based in Warminster, Pa. He has more than 15 years experience in the design and marketing of flowmetering equipment. Mr. Livelli earned an MBA from Regis University and a bachelor’s degree in Mechanical Engineering from New Jersey Institute of Technology. Mr. Livelli can be reached at greg.livelli@us.abb.com or 215 674-6641.

www.abb.com