Recent advances in flow sensing have resulted in more accurate, durable, and economical meters. Increasingly important is the ability to interface meters with a computer for instantaneous flow readout to remotely control flow, or to allow unattended process operation. To meet such demands, new types of flowmeters are being introduced, and older designs are being improved and updated.

Prime considerations when selecting a flow sensor include: the type of fluid being measured, its temperature and pressure, viscosity, conductivity, corrosiveness, and cleanliness. Equally important are the requirements of the sensor itself: Flow velocity range, accuracy, ease of installation, and maintenance requirements. Finally, cost can be a major factor in the decision.

Two basic classes of flowmeters are differential producers and linear flowmeters. Differential-producer flowmeters create a restriction in the flow field. When flow is contracted, either gradually or abruptly, kinetic energy increases at the expense of potential energy (static pressure). The difference between pressure at the full pipe section and that in the vicinity of the contraction is related to the square root of the velocity at the full section minus the square root of the velocity at the contraction. Fluid properties and the abruptness of the contraction also play a role in the operation of these meters. General range is 4:1.

The pressure/flow relationship depends on the length and condition of the reference piping, pressure tap locations, and the geometry of the restricting element (differential producer). Any change in these characteristics alters the relationship, making these devices extremely sensitive to installation conditions. The greatest disadvantage to this class of flowmeters is that they require a secondary measuring system -- flow is determined based on a pressure-flow relationship.

Orifice-plate flow sensors are the most widely used of the differential pressure flow sensors; others being nozzle, venturi, and flow tube. Generally, an orifice sensor is a circular plate inserted between flanges of a pipe with a round or other shape hole bored in the plate center, matching system flow range and characteristics. Pressure taps on each side of the orifice plate measure the pressure differential, and a transmitter produces a signal that is proportional to the square of flow rate. Because the relationship between flow and pressure involves a square root, the differential drops off quickly as flow decreases. For this reason, the dynamic range for these sensors is limited to about 4:1. Accuracy also varies with flow rate, from 1 to 3%.

Venturi tubes produce differential pressure through a section of pipe with a tapered inlet and diverging outlet. The contoured surfaces allow measurement of dirty gas and liquid. Line sizes are typically 2 in. and larger, and accuracy ±1 to 2%. However, the devices are costly.

Flow nozzles are generally used in steam/vapor flows at high velocities. Differential pressure is produced by a pipe section with an elliptical entrance and nozzle exit. In general, line sizes are 2 in. greater, and accuracy is ±1 to 2%. Flow nozzles are expensive and may cause permanent pressure loss.

Pitot tubes provide basic flow measurement for laboratory tasks. A pitot tube consists of a special tube facing into the flow to measure a velocity-augmented impact pressure and a second tube that measures static pressure. Difference between the two pressures is normally measured on a U-tube manometer. A pitot tube measures velocity at only one point in the fluid stream. Accordingly, the sensor must be moved around in the stream so that a variety of readings can be taken to yield a velocity profile.

The annular orifice, consisting of a disk supported concentrically in a pipe section, was developed to overcome the problem of dirt buildup in front of a standard orifice. Operating principle is the same as with a standard orifice meter. It works with both clean or dirty gases and liquids, generally 4 in. and larger line sizes; accuracy is ±2%.

Volumetric flowmeters whose output is not proportional to the square root function are termed linear flowmeters. Meters are either linear due to the principle of operation or linearized through electronic means. Typical range of this class of flowmeters is 10:1.

Turbine flowmeters use a turbine, positioned in a tube, which rotates with passing fluid flow. A proximity sensor detects turbine blade movement and generates a frequency signal that is analogous to flow rate. These sensors are compatible with a variety of fluids, and provide a wide flow range. The dynamic range or turndown ratio (that is, the ratio of maximum to minimum flow rates) is up to 35:1 in some turbine flow sensors. Accuracy over this range is generally about 1%, and response is linear over the entire range, simplifying electronics. Pressure drop across the turbine is quite low (generally about 20 psi at maximum flow), which is an advantage turbine sensors hold over some other types.

Positive-displacement meters offer the ultimate in volumetric accuracy, but require maximum interference with the flow stream. Basically, they function like a hydraulic or pneumatic motor, with the "output shaft" of the motorlike device driving the gage readout. Because these meters are very similar to the hydraulic and pneumatic motors used in fluid-power applications, they are suitable for extremely high pressures. A big advantage of positive-displacement types is their ability to discern extremely low flow, down to a few cc per minute. They are also highly accurate (typically 0.5 to 1% of flow rate), have a dynamic range up to 400:1, and are bidirectional. Hydraulic pulsation has no effect on these sensors, and they can be placed almost anywhere in the system.

Rotameters are the most common variable-area meters, consisting of a tapered tube in which a float is supported by the fluid flowing up through the tube. As the fluid flow increases, the float is lifted higher in the tapered tube, as greater orifice area is required around the float to transmit the flowing fluid. The reading is taken directly from the float position. Rotameters are typically used for low-pressure flow readings; they are most suitable when used for a single fluid.

Vortex-shedding sensors detect flow from viscosity-related effects of a blunt object in a flow stream. Basically, when fluid flows around an object, vortices are shed alternately from one side of the object, and then the other, in a regular pattern. Sensing the rate of vortex passage gives a measure of flow velocity. A number of methods have been developed to detect vortex passage. One is a piezoelectric crystal element that senses induced strain in the shedder bar; another is diaphragm pressure sensors located just beyond the shedder bar. Both have good dynamic range (20:1) and excellent accuracy (0.8%).

Fluid-oscillator meters are based on the Coanda effect, in which a fluid jet adheres to the walls of a Venturi nozzle. When the jet attaches itself to the lower wall, it encounters a flow diverter that splits the flow toward the upper wall. At the upper wall, it encounters a second flow diverter that splits flow back to the lower wall. Flow oscillation is a linear function of flow rate. A heated thermistor placed in the upper feedback passage measures oscillation rate, and hence flow.

Jet-deflection meters use measured flow to detect a high-speed jet from receiving ports. Flow is proportional to the amount of deflection. In some models, the nozzle and port assembly can be moved to obtain a complete velocity-profile distribution curve.

Ultrasonic meters use one of two methods, time-in-flight or Doppler. In the time-in-flight method, used with clean liquids, a high-frequency pressure wave is transmitted across the pipe at an acute angle. The time required for the beam to cross the pipe relates to flow rate. In the Doppler method, used with dirty liquids, the pressure wave is refracted back to a detector by particulates in the fluid. The difference between reflected frequency and transmitted frequency relates to flow rate. The meters are expensive, and accuracy ranges from ±1 to 4%, depending on particulate concentration.

Magnetic flow sensors operate on the principle that the movement of a conductor through a magnetic field induces a voltage. They consist of a flow tube which generates the magnetic field in the pipeline, and an electronic converter that measures induced voltage. The prime limitation of magmeters is that they require a conductive fluid. (Most fluids are, with the notable exception of petroleum-based fluids.) Because there is no obstruction in the pipeline, magmeters have no pressure drop and no parts that will wear out. Solids and contaminants are not a problem, nor is viscosity. These meters are highly accurate, typically 0.5% of flow rate over a dynamic range up to 300:1. But to maintain their accuracy, they must have a flow rate above 1.5 fps.

Magnetic piston flowmeters feature only one moving part, and accurately measure flow from 10 cc/min to 3.5 gpm, at pressures to 3,000 psi. Located in the flow path is a piston-shaped magnet, generally encapsulated in Teflon, that is free to travel in the flow-path bore. A second magnet of opposite polarity is located external to the flow stream, in line with the piston. The resulting magnetic repulsion opposes piston movement and provides resistance to flow.

In operation, flow lifts the piston off its seat, and the piston rises or falls as flow increases or decreases. Adjacent to the flow path, a Hall-effect transducer senses the resulting magnetic field and converts it to a millivolt signal. Because the magnetic field changes with piston position, the voltage produced by the transducer can be directly related to flow.

Laser-Doppler flow sensors differ from other types in that they measure flow at a point, not over an entire area. They are typically used to scan a flow field to gain specific details. The principle of operation involves crossing two laser beams in the flow path, creating interference fringe patterns with alternating light and dark areas. When flow passes through this pattern, light reflected by particles in the fluid correspondingly follows the light density variations, cyclically brighter and darker. The frequency of this cyclic variation is directly related to flow velocity. Dynamic range is about 100,000:1. Typically, these devices can measure velocities in the mm/sec range at the low end, and a few thousand m/sec at the high end. Accuracy in the range of a fraction of 1% is typical, and they are very good at following high-speed transients.