Ultrasonic Flowmeter

What is an Ultrasonic Flowmeter

An ultrasonic flowmeter is a flow-measuring instrument that infers the velocity of a liquid or gas from the way high-frequency sound (typically 0.5 to 4 MHz) propagates through it. Flow, like pressure, temperature, and level, is one of the four fundamental process variables, and the ultrasonic principle has become one of the fastest-growing ways to measure it because the sensing path can be entirely acoustic. The instrument computes volumetric flow rate by multiplying the measured average velocity by the known cross-sectional area of the pipe, so an accurate result depends on both a good velocity measurement and a well-characterized bore.

Functionally, every ultrasonic flowmeter has three parts: one or more transducer pairs that act alternately as transmitter and receiver, an electronics unit that fires the pulses and times their arrival to nanosecond resolution, and a converter that turns velocity into a scaled output (4-20 mA, pulse, Modbus, HART, or a fieldbus). What separates a good meter from a poor one is the timing electronics and the signal processing that rejects noise, multipath echoes, and profile distortion. The transducer is only a piezoelectric element; the intelligence is in how the travel times are measured and integrated.

The transit-time idea dates to the 1950s and 1960s, but early analog electronics could not resolve the sub-microsecond time differences reliably. Practical industrial meters arrived once digital signal processing matured in the 1970s and 1980s. Multipath gas meters for custody transfer became established in the 1990s, and clamp-on meters then opened a separate mass market for non-invasive measurement of water, district heating, and bulk liquids. Today the technology covers everything from a smart residential water meter to a fiscal natural-gas line carrying millions of cubic metres per day.

The defining engineering advantage is that the meter does not contact or obstruct the flow. There is no orifice to create a permanent pressure loss, no rotor to wear out, and no parts to foul in dirty service. A clamp-on meter can even be installed and removed without breaking into the pipe, which makes it the standard tool for surveying large water mains, verifying other meters, and temporary measurement. The trade-offs are sensitivity to the flow profile, dependence on a clean acoustic path, and a higher entry cost than a simple mechanical meter.

Four engineering metrics decide whether a given ultrasonic meter fits a duty: accuracy class (and whether it is quoted as percent of reading or percent of full scale), the fluid cleanliness it can tolerate, the available straight pipe run, and whether the application is fiscal (custody transfer) or merely control or monitoring. Get those four right and the rest of the spec sheet follows; get them wrong and an otherwise excellent meter will read badly in service.

It is worth being precise about what the meter actually measures. The acoustic measurement yields a path-average velocity along a chord, not the true volumetric mean across the whole cross-section. The converter then applies a profile-correction factor, derived from the Reynolds number, to translate chord velocity into mean velocity and finally into volume flow. This is why an ultrasonic meter is described as an inferential volumetric meter rather than a direct mass meter: it does not weigh the fluid the way a Coriolis meter does, and to report mass flow it must be combined with a separately measured or assumed density. For gas custody transfer the density and compressibility are supplied by a flow computer that takes pressure and temperature inputs alongside the meter, following standards such as AGA 8 for gas property calculation.

Types and Construction

Ultrasonic flowmeters are classified two ways at once: by mounting (clamp-on, inline spool-piece, insertion, or hot-tap) and by the number of acoustic paths (single-path versus multipath). These two axes are independent, and the right combination depends on whether the meter must be non-invasive, how accurate it must be, and whether the pipe can be cut. The table below compares the four mounting styles on the criteria that matter to a buyer.

Clamp-on meters strap a pair of transducers onto the outside of an existing pipe; the ultrasonic beam refracts through the pipe wall and any liner before reaching the fluid. Their unique strength is non-invasive installation on DN15 to DN6000 (about 0.5 to 240 inch) pipe with no process shutdown and no pressure drop, which is why they dominate large water mains and retrofit projects. The price is lower absolute accuracy: the electronics must be told the exact wall thickness, liner, and sound speed, and any of those inputs being wrong shifts the reading. Coupling-pad and grease quality also affect long-term signal strength.

Inline (spool-piece) meters build the transducers into a calibrated meter body of known internal diameter, so beam geometry and bore are fixed and traceable. This is the configuration used for permanent process and custody-transfer service, reaching 0.5 percent of reading as standard and better with wet calibration. The cost is that the pipe must be cut to insert the spool, and the wetted transducers must be chemically compatible with the medium.

Insertion and hot-tap meters push a single probe through a pipe fitting into the bore, sometimes under pressure through a valve. They are economical on very large lines where a full spool is impractical, but a single insertion point samples a limited region of the profile and accuracy is correspondingly modest. On the path-count axis, single-path meters sample one chord and assume a known profile, while multipath meters fire several chords at different heights and integrate them. Four to eight paths is typical for liquid custody transfer and five to twelve for fiscal gas, because more paths make the computed mean velocity far less sensitive to swirl and asymmetry.

The path count also buys redundancy and self-diagnosis, not just accuracy. A multipath meter compares the velocities and the measured sound speed on each chord; if one path degrades because a transducer fouls or fails, the meter can flag it, weight it out, and keep metering on the remaining paths within a known reduced accuracy. This is one reason custody-transfer operators favour multipath designs: the meter can prove its own health between calibrations and continue to deliver a defensible measurement after a single-path fault, which a single-path meter cannot. The same diagnostics let a clamp-on survey crew confirm in minutes whether a candidate pipe will give a trustworthy reading before committing to a permanent installation.

Transit-Time and Doppler Principles

Two opposite physical methods both go by the name ultrasonic. Transit-time (time-of-flight) is the high-accuracy method for clean fluids; Doppler is the rugged method for dirty or aerated liquids. They are not interchangeable, and choosing the wrong principle for the fluid is the single most common ultrasonic selection error. The table compares them on the points that determine which one the fluid will allow.

Transit-time sends an ultrasonic pulse diagonally downstream and another upstream along the same path. Travelling with the flow is faster than travelling against it, so the downstream pulse arrives sooner. The transit-time difference (the upstream time minus the downstream time) is proportional to the average flow velocity along the path. In compact form the velocity is V = (delta-t multiplied by c squared) divided by (2 multiplied by L multiplied by cos alpha), where c is the speed of sound in the fluid, L is the path length between transducers, and alpha is the angle of the beam to the pipe axis. A key engineering point is that good designs cancel the dependence on c by combining the upstream and downstream times, so the reading is largely independent of changing fluid sound speed, temperature, and density.

Because transit-time relies on the pulse reaching the far transducer cleanly, it requires a clean, single-phase medium. Suspended solids and gas bubbles scatter and attenuate the beam, which is why transit-time is specified for clean liquids and gases and is generally not recommended for slurry, raw sewage, or wet steam. At zero flow the upstream and downstream times are equal, so a stable, well-calibrated zero is essential; high-end meters measure the times to nanosecond resolution to resolve the small difference at low velocity.

Doppler inverts the requirement. A single transducer (or pair) injects a continuous or pulsed beam, and particles or bubbles carried by the flow reflect it back at a shifted frequency. The frequency shift is proportional to the reflector velocity, and therefore to the flow. Doppler needs reflectors to work at all, so it is the right choice for slurry, raw wastewater, and aerated liquids that defeat transit-time meters. Its accuracy is lower, commonly 1 to 5 percent of full scale, because the reading depends on where in the profile the reflecting particles happen to be and on their concentration.

Some clamp-on transmitters host both methods and switch automatically as fluid cleanliness changes, which is useful across a wastewater plant where one line may run clean and another dirty. For gas, the picture is simpler: practically all industrial gas ultrasonic meters are transit-time multipath designs, because gas carries few reflectors and fiscal accuracy is the whole point. Doppler is essentially a liquid-only technique.

A practical consequence of the transit-time equation is the role of the wedge angle and transducer spacing in clamp-on installs. The transducers sit on plastic wedges that refract the beam into the pipe at a designed angle; the electronics calculates the correct centre-to-centre spacing from the entered pipe diameter, wall thickness, liner, and the sound speed of the fluid. If any of those inputs is wrong, the beam misses the opposite transducer or arrives at the wrong point on the bore, and the reading shifts even though the meter still reports a signal. This is the root of most clamp-on commissioning problems and is why reputable meters display a received-signal-strength and signal-quality figure during setup: a strong, stable signal confirms the geometry is right before any flow number is trusted.

The two methods also behave differently when conditions degrade. A transit-time meter loses signal entirely if bubbles or solids rise above its tolerance, so it fails to a clear no-signal alarm rather than a quietly wrong number, which is desirable for fiscal duty. A Doppler meter, by contrast, keeps producing a reading whose error grows with changing particle concentration and profile, so it can drift without alarming. For this reason transit-time is preferred wherever the consequence of a silently wrong reading is high, and Doppler is reserved for duties where any reading is better than none and the fluid simply will not pass a transit-time beam.

Media, Pipe, and Standards

Three factors outside the electronics decide whether an ultrasonic meter will read correctly: the fluid (does it pass or reflect sound), the pipe (size, material, lining, and straight run), and the standard the measurement must satisfy. Get any of these wrong and the published accuracy never materializes in the field.

Fluid suitability is the first gate. Transit-time meters want a clean, acoustically transparent, single-phase medium: potable and process water, condensate, light and heavy hydrocarbons, natural gas, and many chemicals all work well. They struggle with high solids, persistent gas bubbles, and wet steam. Doppler meters want the opposite, needing suspended particles or bubbles to reflect from. Sound speed in the medium also matters because it sets the beam refraction angle in clamp-on installs; the electronics carries a fluid library so the operator can select water, oil, or a custom sound speed.

Pipe and installation determine the real accuracy more than the brochure does. For clamp-on meters the pipe material and any internal lining must transmit sound: carbon steel, stainless, copper, ductile iron, and many plastics are fine, but porous concrete, heavily corroded walls, or thick mortar lining attenuate the signal. Straight run is critical because the meter samples only a few chords: a common single-path rule is 10 to 15 diameters upstream and 5 downstream, relaxing to as little as 2 diameters upstream for multipath meters with profile correction or an integral flow conditioner. Where straight run is short, fit a tube-bundle or perforated-plate conditioner.

Standards define what accuracy and test regime a meter must meet, and they differ sharply between liquid and gas, and between process and fiscal duty. The table below maps the major designations to their scope.

For fiscal natural gas, a meter is typically expected to satisfy both AGA Report No. 9 and ISO 17089-1, with an OIML R137 accuracy class for legal trade. For billed water, the MID MI-001 module applies. For district heating energy, EN 1434 governs the meter plus paired temperature sensors. Always confirm the certificate covers the exact model, size, and pressure rating you intend to buy, because approvals are granted per configuration, not per brand.

The accuracy classes carry concrete numbers worth knowing. OIML R137 Class 0.5 limits the maximum permissible error to roughly 0.5 percent above the transition flow rate and about 1 percent at the lowest flows, while Class 1.0 doubles those limits; published custody gas meters such as the FLOWSIC600-XT and ALTOSONIC V12 are certified to these classes. AGA Report No. 9 sets analogous flow-weighted accuracy and repeatability limits for multipath gas meters and prescribes a flow-calibration regime. Crucially, these standards specify not only the meter error band but the calibration and proving method, the installation effects to be tested, and the diagnostics the meter must report, so a meter that merely claims a percentage without the standard behind it is not comparable to a certified one.

Key Specification Parameters

A datasheet may list 20 or more lines, but eight parameters drive the selection decision: accuracy and its reference basis, repeatability, velocity range and turndown, nominal diameter range, medium and ambient temperature, number of acoustic paths, output signals, and hazardous-area and metrology approvals. Each is explained below.

Accuracy and its basis is the most misread line. Ultrasonic meters are usually quoted as percent of reading (o.r.), which holds the error proportional across the range, unlike percent of full scale (FS), where the error is fixed and dominates at low flow. A good inline transit-time liquid meter reads 0.5 percent o.r. as standard, dropping to 0.15 to 0.3 percent on multipath custody designs after wet calibration; clamp-on is typically 0.5 to 2 percent o.r. depending on pipe data quality. Doppler is normally 1 to 5 percent FS. Never compare an o.r. number against an FS number directly; convert both at your real operating point.

Repeatability is the scatter of repeated readings at the same flow and is often better than accuracy; multipath meters reach 0.1 to 0.2 percent. It is the figure that matters for totalization and batch repeatability even when absolute accuracy is modest. Velocity range and turndown follow from the meter having no moving part to set a minimum: usable velocity is roughly 0.01 to 12 m/s for liquids, giving turndown of 100:1 to over 250:1. Accuracy still falls near zero velocity, so size so normal flow exceeds about 0.3 m/s.

Nominal diameter is where ultrasonic outranges every other technology: clamp-on liquid meters cover DN15 to DN6000 (0.5 to 240 inch) from one transmitter with swappable transducers, and inline spools run from about DN15 to DN3000. Temperature has two limits: the medium temperature the wetted parts or coupling pads tolerate (clamp-on coupling pads commonly span about -40 to +200 degrees C across product families, with specific sets narrower) and the ambient temperature for the electronics. Confirm both against the datasheet for the exact transducer set.

Output signal is the interface to the control system. The mainstream options are:

• 4-20 mA (with HART): the default analog output for flow rate, with HART for remote configuration and diagnostics.
• Pulse / frequency / status: scaled volume pulses for totalizers and batch controllers, plus alarm status contacts.
• Modbus RTU / TCP: digital register access for SCADA and data logging, common on clamp-on transmitters.
• Fieldbus and industrial Ethernet: PROFIBUS PA, FOUNDATION Fieldbus, PROFINET, and EtherNet/IP for large DCS projects.
• Bidirectional flow and diagnostics: forward and reverse totals plus signal-strength, sound-speed, and gain diagnostics that flag a degrading acoustic path.

Number of paths sets both accuracy and immunity to flow disturbance, as covered in Chapter 2. Approvals cover hazardous area (ATEX, IECEx, FM, NEPSI) for flammable media and legal metrology (OIML R137, MID, AGA 9 acceptance) for fiscal duty. A meter without the right certificate cannot be used for trade no matter how accurate it is.

Selection Decision Factors

To turn the previous chapters into a specific model, follow the decision sequence below. Most ultrasonic selection mistakes come not from one wrong line on a datasheet but from deciding the meter style before the fluid and the duty are settled. These eight steps double as an RFQ template.

1. Fluid and cleanliness: First confirm the medium is clean single-phase (transit-time) or particle-laden / aerated (Doppler). This single answer eliminates half the catalog and prevents the most common error of all.
2. Duty class: Decide whether the measurement is fiscal / custody transfer (needs OIML R137, AGA 9, ISO 17089-1, or MID), process control (0.5 percent o.r. inline is usually enough), or simple monitoring (clamp-on at 1 to 2 percent o.r.). Each tier roughly doubles cost.
3. Mounting and pipe access: Choose clamp-on if the pipe cannot be cut or is very large, inline if accuracy and permanence dominate, insertion / hot-tap for huge lines under pressure. Record exact pipe material, wall thickness, and lining for clamp-on.
4. Diameter and velocity: Match DN to the meter range and check that normal flow sits above about 0.3 m/s and below roughly 12 m/s. For wide-ranging flow, exploit the meter high turndown rather than oversizing.
5. Straight run and conditioning: Verify available upstream and downstream diameters against the datasheet; if short, add a flow conditioner or move to a multipath meter with profile correction.
6. Temperature and pressure: Check medium temperature against transducer or coupling-pad limits and ambient against the electronics rating; confirm pressure rating of any spool or hot-tap fitting.
7. Output and integration: Match 4-20 mA / HART, pulse, Modbus, or fieldbus to the control system, and specify bidirectional totals and diagnostics where flow can reverse or the path can foul.
8. Certifications and total cost: Confirm hazardous-area and metrology approvals cover the exact model and size, then compare total cost of ownership: clamp-on saves installation and shutdown cost, inline custody meters cost more but earn it back in measurement uncertainty on high-value flows.

One last dimension is commonly overlooked: serviceability and diagnostics. Ultrasonic meters expose rich health data, including signal strength, gain, measured sound speed, and per-path validity, which let a maintenance team detect a fouling path or a drifting coupling pad before the reading goes wrong. Confirm the meter logs these diagnostics and that the vendor offers local calibration, transducer spares, and firmware support. KROHNE, Endress+Hauser, Emerson Daniel, Siemens, Honeywell Elster, and Baker Hughes maintain calibration and service capability for fiscal meters, while clamp-on suppliers such as Flexim, Katronic, and Dwyer/Omega support retrofit and portable fleets.