What accuracy means in flow measurement
Catalog accuracy on a flow meter runs from ±0.1 % on a high-accuracy Coriolis to ±1.5 % on a metal-tube rotameter, with vortex meters in the middle at ±0.5 %. The number is the deviation under a very specific set of reference conditions: clean fluid, calibrated factory rig, transmitter zero confirmed, 10×DN straight pipe upstream, temperature and pressure inside the calibration range. Most field readings sit outside one or more of those conditions, so the catalog spec and the installed accuracy are two different numbers.
What I see on quote requests: buyers paste a catalog accuracy number into the contract without asking what conditions the calibration was witnessed under. The first call after start-up is usually about an installed reading 1-2 % off the catalog number. The catalog was never wrong. The installation sat outside the calibration range.
Jacky Wu, Senior Application Engineer
Catalog accuracy is published in one of two anchor formats. The format choice isn’t cosmetic; it controls the actual error across the working range, and a buyer who reads one as the other will mis-size the meter by an order of magnitude.
% of Rate (% R)
Error scales with the actual flow reading. A ±0.5 % R meter reading 100 m³/h has a ±0.5 m³/h error; the same meter reading 10 m³/h has ±0.05 m³/h. Format used by every rate-based principle: turbine, vortex, Coriolis, EMF, ultrasonic, swirl, oval gear, gas Roots.
Wins on high-turndown variable-load duty because the error scales with the flow.
% of Full Scale (% FS)
Error fixed at maximum flow rated. A ±1 % FS meter rated for 100 m³/h has a fixed ±1 m³/h error, regardless of whether the actual flow is 100 m³/h or 10 m³/h. Format used by thermal mass (sometimes), variable-area rotameters, and differential-pressure (DP) primaries at low turndown.
Fits steady fixed-load duty. At 10 % of FS the error consumes 10 % of the reading. That gap is the format’s structural limit.
The math at 10 % of rated maximum flow:
±1 % R = ±1 % × 0.1 × Qmax = ±1 % of actual reading A ten-fold gap in the realized error at the low end of a 10:1 turndown, for the same nominal catalog accuracy number. The gap widens further at 20:1, 50:1, 100:1 turndown.
Mixing the two formats in a comparison isn’t a minor convention error—it’s the single most common spec-sheet mistake.
A vortex at ±0.5 % R and a rotameter at ±1.5 % FS look nominally close at full scale. At 20 % of FS the gap opens up:
- Vortex (% R): ±0.5 % of reading, unchanged
- Rotameter (% FS): ±7.5 % of reading, fifteen times wider
Custody-transfer specifications always anchor on % of rate. Check-meter and visual-indication duty can live with % FS because the reading is checked against a procedure-fixed flow.
The conversion rule between the two formats isn’t algebraic. They describe different physical guarantees. A % FS spec can be re-expressed as the % R value at a specific flow, but the % R format implies a behavior across the entire working range that % FS doesn’t. Always specify the format explicitly on a purchase order.
Primary-element accuracy versus system accuracy
Catalog accuracy describes the primary element on its own. System accuracy adds three more error sources: the transmitter, the installed piping, and the fluid properties. Three cases:
DP primary + DP transmitter. The DP primary (orifice / venturi / nozzle / wedge / pitot) is rated to ±0.5 % of flow under ISO 5167. The transmitter (Rosemount 3051S, Yokogawa EJX110, Honeywell ST3000) adds ±0.075 % of full pressure range. At full pressure differential the transmitter error is small and system accuracy stays near ±0.5 % R. At 1/16 of full differential (bottom of a 4:1 flow turndown, since Q ∝ √ΔP) the transmitter error dominates the working signal and system accuracy degrades to ±1.5 to ±2 % of rate. Custody-grade DP installations stage two or three transmitters across the range so each one operates near its rated maximum.
Coriolis. The Coriolis primary is the resonant tube. It reads mass directly from the Coriolis force on the moving fluid, so system accuracy equals primary accuracy. No separate density correction. No transmitter compounding. No upstream straight-pipe sensitivity. Coriolis ZLMFM Series publishes ±0.1 % of rate on liquid at the top tier, with five accuracy tiers running ±0.1 % to ±1.0 % of rate.
Ultrasonic. The ultrasonic primary reads transit-time difference between upstream and downstream acoustic paths. System accuracy depends on beam-path geometry, wall-thickness assumption, fluid sound-speed assumption, and the number of paths. A single-path clamp-on holds ±2.0 % of rate; a single-path inline spool-piece (high-tier X8 / X7) holds ±0.5 % at the same nominal velocity. The path geometry is the system.
The general rule: treat catalog accuracy as the floor of achievable system accuracy. Ask the vendor four things before signing: what the spec includes, what it excludes, what calibration conditions were used, and what installed deviation to expect.
Accuracy grades
Four grades:
- Check-meter / indication grade: ±1.5-2.5 % typical. Local visual confirmation. Variable-area rotameters sit here. No traceable calibration required beyond factory cert.
- Industrial grade: ±0.5-1.0 % of rate. Process control, energy accounting, internal audit. Most vortex, EMF, turbine, thermal mass, gas Roots, oval gear, swirl meters. Traceable wet-cal to a national standard. Fiscal use not authorized.
- Custody-transfer grade: ±0.1-0.5 % of rate. Coriolis at ±0.1 %, high-tier ultrasonic ±0.5 %, oval gear at ±0.2 %, liquid turbine at ±0.5 %. Requires a notified-body approval certificate under OIML R 117 / R 137, MID MI-002 / MI-005, API MPMS 5.2, or AGA-3 / AGA-9. Vendor “claims R 117 conformance” without the cert is reference text, not approval.
- Fiscal grade: accounting and tax duty. Frequently overlaps custody-grade. Adds site-acceptance test, periodic recalibration, sealed transmitter, anti-tamper documentation.
The honest claim for a factory-direct meter names the standard and the application: “designed per OIML R 117 conventions used in custody practice” — never “approved” unless a certificate from a notified body ships with the unit. Four reference standards cover most service:
- ISO/IEC Guide 98-3 (GUM): uncertainty budgeting.
- ISO 5167: DP primary elements.
- OIML R 117: dynamic liquid measuring systems.
- NIST HB-44 §3.30 / §3.36: US field accuracy classes.
Accuracy by instrument principle
Catalog accuracy by principle. Read each row as primary-element accuracy under reference conditions; system accuracy compounds per the rules above.
| Series | Anchor format | Typical accuracy | Notes / catalog basis |
|---|---|---|---|
| Coriolis | % of rate | ±0.1 % liquid mass / ±0.5 % gas / ±0.5 % volume | Five tiers ±0.1 / ±0.15 / ±0.2 / ±0.5 / ±1.0 % R on ZLMFM. Mass-direct, system = primary. Handles universal media: HFO, cryogenic, 2-phase. Cost premium 2-3× over velocity meters. |
| DP primary (orifice / venturi / nozzle / pitot) | % of flow | ±0.5 % primary + ±0.075 % transmitter | Per LG primary-element Series. System accuracy compounds with DP transmitter (±0.075 % of full pressure reading) and degrades at bottom of √-extracted turndown. ISO 5167-conformant install assumed. |
| Vortex (LUGB Series) | % of rate | ±0.5 % R liquid · ±1.0 % R gas/steam (typical) | ±0.5 / ±1.0 / ±1.5 / ±2.5 % per LUGB. Built-in T+P on gas/steam. 1:70 turndown at ±0.5 % on liquid per VRd 2024 LUGB award. |
| Ultrasonic (transit-time) | % of rate | ±0.5 to ±2.0 % single-path · ±0.5 % multi-channel inline | Clamp-on (X3/X3M) ±2.0 %, inline spool-piece (U7/U8) ±0.5 % single-path, 2- and 4-channel inline (D8 series) ±0.5 %. Path-count and beam geometry are part of system accuracy. |
| Electromagnetic (LEA Series) | % of rate | ±0.5 % R standard · ±0.2 % R high-accuracy | ±0.5 % R code A, ±0.2 % R code B per LEA. Loosens to ±1.0 % at DN ≤ 25. Conductive liquid only (≥ 5 µS/cm). |
| Liquid turbine (LWGY Series) | % of rate | ±0.5 % R standard · ±1.0 % R / ±2.5 % R economy | ±0.5 / ±1.0 / ±2.5 % per LWGY code A/B/C. Repeatability ±0.1 %. Viscosity ≤ 10 cSt. K-factor drifts 0.1-0.3 %/year clean, faster on particulate. |
| Gas turbine (LWGQ Series) | % of rate | ±0.5 % R high-grade · ±1.0 / ±1.5 % R standard | Catalog accuracy ±0.5 / ±1.0 / ±1.5 % per the LWGQ accuracy table. Built-in T+P compensation. Dry gas only (no condensing wet gas, no 2-phase). |
| Gas Roots (LQ Series, PD) | % of rate | ±0.5 % R standard · ±1.0 / ±1.5 % R tiers | ±0.5 / ±1.0 / ±1.5 % per LQ code A/B. Repeatability ±0.1 %. PD physics holds across full turndown including low-flow billing. |
| Thermal mass (TMF Series) | % of reading (typical) / % FS in some builds | ±1.0 % reading grade A · ±1.5 % grade B | ±1.0 % grade A, ±1.5 % grade B per TMF. Calibration-gas pair-specific; mismatch drifts the reading 1-5 %. |
| Variable area (LZD metal-tube rotameter) | % of full scale | ±1.5 % FS standard · ±2.5 % FS economy | ±1.5 % FS code A, ±2.5 % FS code B per LZD. Local visual indication, not custody-grade. Reading drift at bottom 20 % is structural. |
| Oval gear (PD, viscous liquid) | % of rate | ±0.5 % R standard · ±0.2 % R precision tier | ±0.5 % R base, ±0.2 % R precision tier (suffix −J). Repeatability ±0.05-0.1 %. Volumetric-PD physics: accuracy improves at higher viscosity. |
| Swirl (LUXB Series) | % of rate | ±0.5 % R high-grade · ±1.0 / ±1.5 % R tiers | ±0.5 / ±1.0 / ±1.5 % R per LUXB. Built-in P+T+compressibility on gas custody. 3D upstream / 2D downstream straight-pipe; tightest install allowance on site. |
The conversation I have with most buyers comparing this table to a Western-OEM datasheet: the headline accuracy is the same, the price isn’t. The cost difference on the brand-name unit pays for the traceability chain to PTB or NIST, not a tighter physical measurement. If the audit doesn’t need a PTB stamp, the same accuracy class with a NIM-traceable cert costs 50-60 % less. I write this so buyers stop confusing the cert with the meter.
Jacky Wu, Senior Application Engineer
Frequently asked questions
What does ±0.5% accuracy actually mean on a flow meter?
Because the absolute error on a % FS spec is fixed at maximum flow, while % R scales with actual flow.
At 10 % of full scale:
- ±1 % FS meter delivers ±10 % of rate
- ±0.5 % R meter delivers ±0.5 % R
A twentyfold gap for the same nominal catalog number. Variable-load duties always specify % R: compressed air mains, burner headers, batch dosing, seasonal billing meters. Steady fixed-load duties can live with % FS: pump discharge at design point, constant chiller header.
How do I calculate flow meter accuracy on an installed system?
Catalog accuracy describes the primary element under reference conditions. System accuracy compounds the primary error with installed contributors:
- DP transmitter zero drift
- Turbine K-factor calibration drift
- Thermal-mass calibration-gas mismatch
- Installed velocity-profile distortion
Worked case: a ±0.5 % of flow primary paired with a ±0.075 % transmitter. At 4:1 turndown the system compounds to ±0.6 to ±0.9 % R at the bottom of the range per ISO/IEC Guide 98-3 (GUM).
Match catalog conditions against installed conditions before signing a contract accuracy.
Why does repeatability matter more than absolute accuracy for batch dosing?
The batch-control loop uses the meter as the feedback signal. The operator tunes the recipe target to the meter reading, so the absolute error is absorbed on day one.
The number that matters batch-to-batch is repeatability: how close successive readings of the same flow event come to each other. A turbine at ±0.5 % R accuracy with ±0.05 % R repeatability doses 1,000 kg consistently to ±0.05 kg per batch.
Fiscal billing cares about absolute accuracy. Process control and batch dosing care about repeatability.
Final Thoughts
Reading an accuracy spec correctly is the difference between an installed meter that meets the contract and one that drifts outside it. Two numbers always travel together: the catalog accuracy on the brochure and the calibration uncertainty on the wet-cal cert.
- Anchor on % of rate for variable-load and custody duty; the deviation tracks down with the flow.
- Use % of full scale only on steady fixed-load duty where the reading is checked against a procedure-fixed flow.
Treat catalog accuracy as the floor of system accuracy, then compound it with transmitter drift, installed piping, and fluid-property contributions before signing the contract. The catalog is the conversation starter—the cert is the contract.
Need an accuracy budget for a specific duty? Send pressure, flow turndown, fluid composition, line size, and contract accuracy, and our engineering desk returns a sized quote with the full breakdown (primary + transmitter + install contribution) within four business hours.
Need an accuracy budget for a specific duty?
Send the service data (pressure, flow turndown, fluid composition, line size, contract accuracy) and our engineering desk returns a sized quote with the system accuracy breakdown (primary + transmitter + install contribution) within 4 business hours.