Vortex flowmeters work on the Kármán vortex street principle: a bluff body sheds alternating vortices at a frequency linearly proportional to flow velocity, and the meter counts that frequency to infer volumetric or (with density compensation) mass flow [S5].
Selection is governed less by brand than by whether the line Reynolds number stays above the meter's stated threshold, the fluid stays single-phase, and the temperature/pressure envelope fits the sensor body — three gates that eliminate roughly half of candidate applications before accuracy class is even discussed [S1][S2][S4].
Fluid phase and Reynolds window — the gate that filters half the bids
Vortex meters are designed for turbulent, single-phase flow of liquids, gases, and steam; they fail silently on multiphase or laminar streams because vortex shedding loses amplitude below the meter's minimum Reynolds threshold [S5]. Kaifeng Instrument publishes accuracy of Class 1.0 and 1.5 for gas and Class 0.5 and 1.0 for liquid across DN15–DN300, paired with 24 VDC external or 3.6 VDC lithium-battery internal power, Ex db IIC T4–T6 Gb and Ex ia IIC T4 Ga explosion-proof marks, and IP67 housing [S4]. The dual-certification pattern (flameproof + intrinsic safety) is the practical signal that a vendor has documented both hazardous-area and battery-powered skids for the same SKU. For steam, condensate, and high-temperature gas, ABB's VortexMaster FSV450 documents direct mass and standard-volume outputs by reading remote pressure, temperature, or density on an integrated 4–20 mA interface and combining them with the meter's own flow and temperature signals [S2].
Oval Corporation's Hybrid Multi Delta vortex variant specifies a volumetric flow range of 0.6–13,440 m³/h with process temperature 0–50 °C and process pressure 0–7.8 bar, framing the gas/in-line low-pressure envelope typical of utility and HVAC-class applications [S1]. That ceiling is the reason most steam services above ~10 bar or ~200 °C route to different body/pressure-rating SKUs within the same vendor's vortex family.
Line size, accuracy class, and the comparison that drives the PO
Selection collapses into four decision criteria once fluid phase is fixed: nominal diameter range, accuracy class, output/communication stack, and explosion-proof/ingress rating. A side-by-side view of three currently-published vortex lines shows how those criteria differ in practice. [S1]
Two patterns emerge from the table. First, only Kaifeng's published excerpt gives a fully quantified accuracy class by phase, which is the most defensible spec to put on a data sheet when no vendor selection has been made [S4]. Second, the universal three (ABB, Yokogawa, Kaifeng) target steam and gas/liquid simultaneously, while Oval's Hybrid Multi Delta is a gas/in-line specialty product [S1][S2][S3][S4]. Engineers bidding steam services should weight the FSV450's documented mass-flow compensation path, since standard volume on steam without P/T input typically loses accuracy class.
Who a vortex flowmeter is FOR — and who it is NOT for
Vortex meters are FOR clean, single-phase, turbulent flow of steam, saturated gas, and low-viscosity liquid in lines generally above DN15 and at velocities above the meter's documented minimum [S4][S5]. They are particularly well-matched to saturated and superheated steam mass-metering skids, compressor discharge lines, and boiler-feed condensate where the bluff-body shedding frequency is stable and the line is large enough to develop a Kármán street [S2][S5].
Vortex meters are NOT for slurry, two-phase flow, low-pressure wet gas, or lines that run laminar — the bluff body simply does not shed reliably, and readings drift without a clear error flag [S5]. They are also a poor fit for very low flow rates below the meter's stated minimum Reynolds, where the shed amplitude collapses into the noise floor of the piezoelectric or strain-gauge sensor. For clean-in-place verification of rinse-water quality, a [turbidity meter](/news/turbidity-meter-selection-criteria-for-clean-in-place-verification.html) is a more honest measurement technology than a vortex meter repurposed for a job it cannot resolve. Likewise, custody-transfer hydrocarbon services typically route to Coriolis flowmeters or ultrasonic flowmeters when density-corrected mass and low Reynolds performance are required, leaving vortex meters to the steam and utility envelope where their cost-per-meter-of-line is hard to beat.
Sensor, output, and comms — what the data sheet must say
Yokogawa's digitalYEWFLO VY series explicitly markets a sensor design that "accurately identifies and eliminates noise" so that no startup tuning is needed, a feature that matters on lines with high vibration from adjacent pumps or compressors where a cheaper bluff body would false-trigger [S3].
Output and comms drive the integration cost more than the meter cost. The FSV450 ships with a two-wire 4–20 mA + HART transmitter plus an integrated 4–20 mA input for remote P/T/density, which lets one meter serve as a small flow computer for mass and standard-volume outputs [S2]. Kaifeng's LUGB line is 4–20 mA + pulse with 24 VDC external or 3.6 VDC lithium internal power, the latter being the option that lets a meter be dropped on a remote utility header without running instrument power [S4]. The exploded vortex flowmeter reference on this site covers the Kármán shedding math and the published minimum Reynolds thresholds by body geometry.
Installation, vibration, and the failure modes that show up after commissioning
The single most common field failure on a vortex meter is vibration false-triggering on a line with a pulsating pump or compressor upstream; the sensor counts pipe vibration as vortex shedding and over-reads flow. Yokogawa's noise-isolation DSP sensor targets exactly that failure mode [S3]. The second is two-phase flow — a steam line that carries condensate slugs at turn-on will read incorrectly because the bluff body is shedding in liquid, not vapor, and the calibration constant is wrong for the actual fluid [S5]. The third is low-flow cut-off mis-set: a meter with a Reynolds threshold above the actual line Reynolds will keep reading zero long after flow has started, and the operator logs it as "meter failed."
These are generic good-practice rules — the exact values are published in each vendor's installation manual and should be checked against the chosen SKU before fabrication. For services where vibration is unavoidable, electromagnetic flowmeters on conductive liquid or turbine flowmeters on clean gas are more honest alternatives than a vortex meter that needs a vibration-isolating spool.
Standards, certifications, and how to read the data sheet
Explosion-proof certification is the most visible compliance item on a vortex meter data sheet. Kaifeng's LUGB line carries Ex db IIC T4–T6 Gb (flameproof, hydrogen-safe, up to 135 °C gas temperature class) and Ex ia IIC T4 Ga (intrinsically safe, Zone 0 capable), plus IP67 [S4]. The two marks are not redundant: flameproof covers the cast-body and electronics compartment, intrinsic safety covers the low-energy signal loop. Skids that mix a 24 VDC-powered main meter with a battery-powered remote header meter typically need both marks in the bill of materials.
Other items to verify on the data sheet: wetted material (commonly 316L stainless), flange standard (ASME B16.5, EN 1092-1, JIS B2220 — match the line), pressure rating (PN16/40/63 or 150#/300#/600#), and a published flow range vs Reynolds curve. Mass-flow accuracy on steam and gas only holds when the pressure and temperature inputs are within the meter's stated compensation range, which is why the FSV450's integrated 4–20 mA input for P/T/density is a meaningful spec line rather than a marketing bullet [S2]. Turbine flowmeters remain the more conservative pick for clean, low-viscosity gas services below DN50 where a small K-factor and well-published Reynolds performance are required. For a broader context on open-channel flowmeters and the part-flume / part-weir applications that vortex cannot serve, the encyclopedia entry lays out the side-by-side.
Decision checklist before the PO is cut
Confirm four things in this order before specifying: (1) fluid is single-phase, (2) line Reynolds stays above the meter's published minimum across the turn-down range, (3) process temperature and pressure fall inside the chosen SKU's body rating, (4) output stack matches the DCS/PLC card population. If any of those four fail, the meter is wrong regardless of accuracy class on the data sheet [S1][S2][S4].
Two trackable signals to watch after commissioning: a baseline noise-floor reading on the sensor output at zero flow (anything >2–3% of URL is a vibration flag), and a 30-day comparison of the meter's integrated total against a downstream custody or check meter (drift >0.5% on steam or >1% on gas is a Reynolds- or compensation-input problem). Engineers working on a parallel differential-pressure project can cross-check the operating envelope against the [differential pressure transmitter buying guide](/news/differential-pressure-transmitter-buying-guide-2026-spec-comparison.html) since the impulse-line and manifold design rules overlap.