Understanding ISO 5167: A Practical Guide for Engineers

ISO 5167 is the international reference for DP flow measurement without individual calibration. When your meter run satisfies all applicable requirements — geometry, installation, operating limits, and fluid conditions — you can calculate the discharge coefficient from the standard's equations with a defined uncertainty. This eliminates the cost and schedule impact of wet calibration while providing an auditable, repeatable calculation trail. For custody-transfer applications, ISO 5167 traceability is often written into commercial contracts and government regulations: the meter is acceptable because the standard is accepted. For process control, ISO 5167 compliance provides a defensible engineering basis that survives internal audits, HAZOP reviews, and insurance surveys. The 2022 edition consolidated all six parts into a single coordinated publication with updated Reynolds-number limits, refined uncertainty statements, and new guidance on installation effects. If your internal specification still references ISO 5167-1:2003, update it: the 2022 edition tightened several requirements and your meter may be non-compliant on paper even if it works in practice.

Part 1 is the common foundation. It defines the symbols, terms, uncertainty calculation methodology, and installation requirements that apply to every device covered by Parts 2 through 6. Key provisions every engineer should know: (1) The pipe must run full and the fluid must be single-phase or treated as single-phase with defined corrections. The standard does not apply to multiphase flow. (2) The pipe Reynolds number must be above the minimum specified for each device — if ReD drops below the limit, the Cd equation is no longer valid and calibration is required. (3) The velocity profile at the measurement cross-section must be fully developed and free from swirl. Swirl from two out-of-plane elbows or a partly closed upstream valve cannot be corrected by a longer straight run alone; a flow conditioner is required. (4) Uncertainty is expressed at the 95% confidence level (coverage factor k=2) and combines geometric, coefficient, and installation contributions by root-sum-square. (5) The expansibility factor ε applies to all compressible fluids; Part 1 provides the general method and each device-specific part provides the applicable correlation. (6) Pressure taps must be individually drilled — annular grooves or averaging rings are not covered by ISO 5167 and require separate calibration. (7) The pipe internal diameter must be measured, not taken from nominal pipe tables; a 1 mm error on DN 100 shifts the flow result by approximately 0.5%.

Part 1 also defines the straight-run requirements framework: three columns in the straight-run tables for each device, corresponding to zero additional uncertainty, +0.5% additional uncertainty, and the shortest length at which the supplier has demonstrated acceptable performance. Engineers who always use the zero-additional-uncertainty column are over-specifying for process control and adding unnecessary piping cost. Part 1 gives you permission to use a shorter run — provided you accept a known and quantified uncertainty penalty.

Part 2 covers square-edged concentric orifice plates with corner, flange, and D and D/2 pressure taps. This is the most widely used DP primary element globally and the most extensively validated. Key constraints: beta ratio β = d/D must be between 0.10 and 0.75. The minimum pipe Reynolds number ReD is β-dependent: ReD ≥ 4,000 for β ≤ 0.45, approximately 40,000 at β = 0.60, and approximately 160,000 at β = 0.75. The discharge coefficient Cd is calculated from the Reader-Harris/Gallagher equation, which is a function of β, ReD, pipe diameter D, and tap location. Typical Cd values range from 0.59 to 0.62. The upstream edge must be square and sharp — the edge radius must not exceed 0.0004d. A 0.1 mm radius on a 50 mm bore is non-compliant and will under-read. Plate flatness must be within 0.005(d/β) per 25 mm. The expansibility factor ε follows the empirical ISO 5167-2 correlation: ε = 1 − (0.351 + 0.256β⁴ + 0.93β⁸)[1 − (p2/p1)^(1/κ)]. This is specific to orifice plates — it does not apply to nozzles or Venturi tubes. Apply it to any other element and the expansibility will be wrong.

The straight-run requirements occupy Table 3 of Part 2 and are the most consulted table in DP flow measurement. For β = 0.65 with a single 90° elbow upstream, the required length is 22D for zero additional uncertainty, 16D for +0.5% additional uncertainty. For two elbows in perpendicular planes at the same β, the length jumps to 34D for zero additional uncertainty. A flow conditioner (compliant with ISO 5167-2 Annex B) can reduce this to 5D plus the conditioner length. The most common orifice-plate non-compliance is a flow conditioner that is physically installed but not compliant with the standard — wrong hole pattern, wrong thickness, or wrong location relative to the plate.

Part 3 covers three geometries: the ISA 1932 nozzle, the long-radius nozzle, and the Venturi nozzle. The ISA 1932 nozzle has a smooth, progressively curved inlet profile that guides flow onto the cylindrical throat with minimal separation. It is used primarily in Europe for high-velocity steam and gas applications, and in power stations for feedwater and main steam lines. The long-radius nozzle has a simpler quarter-ellipse inlet and a longer throat; it is more common in North America and handles high-temperature applications well. The Venturi nozzle combines a convergent nozzle inlet with a truncated Venturi diffuser for partial pressure recovery. Key constraints: β limits are 0.30 to 0.80 for ISA 1932, 0.20 to 0.80 for long-radius. The ReD lower limit is 15,000 for ISA 1932 nozzles or 10,000 for long-radius nozzles. Cd is calculated from empirical correlations that are distinct from the orifice-plate equation: ISA 1932 Cd depends on both β and ReD and ranges from approximately 0.95 at ReD = 20,000 to 0.99 at ReD = 2×10⁶.

Critical distinction: the expansibility factor for nozzles is fundamentally different from orifice plates. ISO 5167-3 uses the thermodynamic isentropic expansion formulation, not the empirical orifice-plate ε equation. The isentropic exponent κ must be evaluated at upstream conditions. Using the orifice ε on a nozzle produces a systematic error that increases with ΔP/P — at ΔP/P = 0.15, the error can exceed 0.5%. Another pitfall: the ISA 1932 nozzle requires a downstream pressure tap at a specific location that depends on β; using corner taps (as with an orifice) is non-compliant and will produce incorrect differential pressure. The nozzle's smoother profile also makes it more erosion-resistant than an orifice plate in wet steam and high-velocity gas, which is why you see long-radius nozzles on main steam lines in power plants where replacement during operation is impossible.

Part 4 covers classical Venturi tubes in three construction categories: machined, rough-cast, and fabricated. The classical Venturi consists of a convergent inlet (21° ± 1° included angle), a cylindrical throat, and a divergent diffuser (7° to 15° included angle). The convergent accelerates the fluid smoothly; the throat sustains a stable low-pressure measurement zone; the diffuser decelerates the fluid and recovers 80–90% of the pressure drop generated at the throat. Cd is high: 0.984–0.995 for machined Venturi tubes within their ReD validity range (ReD ≥ 2×10⁵ for machined, ReD from 2×10⁵ to 2×10⁶ for rough-cast). The machined Venturi has the lowest uncertainty of any uncalibrated DP element because its Cd is close to unity and weakly dependent on ReD. β limits are 0.40–0.75 for machined, 0.30–0.75 for fabricated.

The diffuser angle is the most critical manufacturing tolerance. ISO 5167-4 limits it to ≤15°; angles above this cause boundary-layer separation in the diffuser, which destroys pressure recovery and may introduce instability. A 30° diffuser can reduce recovery from 85% to 55%, eliminating the Venturi's primary advantage. In practice, this means classical Venturi tubes are long — 5D to 8D overall length — and heavy. A DN 300 fabricated Venturi with Class 300 flanges can weigh over 500 kg and require structural supports. This is the tradeoff: the best pressure recovery in DP metering, at the cost of the largest physical envelope and highest purchase price. The expansibility factor for Venturi tubes uses the thermodynamic formulation, same as nozzles, because the smooth convergent produces near-isentropic acceleration. The downstream pressure tap must be in the throat, not in the diffuser — measuring at the diffuser exit captures partial recovery and produces an incorrect differential.

Part 5 covers V-Cone and similar cone-type DP meters, where a cone-shaped restriction is mounted concentrically in the pipe and flow passes through the annular gap between the cone and the pipe wall. The defining geometric parameter is the beta ratio: β = √(1 − dc²/D²), where dc is the cone base diameter and D is the pipe diameter. The cone forces the flow to the pipe wall, creating a stable, ring-shaped velocity profile that makes the cone meter less sensitive to upstream disturbances than an orifice plate — tested installations can operate with as little as 3D to 5D upstream straight run. The upstream pressure tap is located before the cone face; the downstream tap is in the cone's wake, typically through the cone support. Key constraints: β range 0.45–0.75, ReD ≥ 8×10⁴ for standardized use. Below ReD = 8×10⁴, individual calibration is required because the Cd correlation has not been validated in the transitional regime. The Cd uncertainty for a standardized cone meter is ±5% without calibration and ±0.5% with calibration, reflecting the element's more complex flow field. Pipe diameter range: D = 50–500 mm.

Critical distinction: β for a cone meter is NOT d/D. Using the orifice-plate formula β = d/D for a cone will produce a radically wrong beta ratio and a completely invalid sizing calculation. The cone β formula β = √(1 − dc²/D²) must be used. Also critical: the cone support structure (pipe-wall supports or a central support strut) creates a wake that affects the downstream pressure tap reading. The tap location is not arbitrary — it is defined by the manufacturer's design and validated by calibration. Moving the tap changes the measured Cd. ISO 5167-5 also specifies limits on the cone angle: the cone half-angle is typically 29° to 35°; angles outside this range may not produce the flow-separation pattern assumed by the standard. For the Shizhong V-Cone, the cone half-angle is 29° with a cone length-to-diameter ratio designed to produce stable separation across the β range.

Part 6 covers wedge meters — a V-shaped restriction mounted in the pipe such that flow passes through a segment-shaped opening at the top or bottom of the pipe cross-section. The defining parameter is the ratio h/D, where h is the height of the opening (gap) above the wedge apex. The beta ratio β is defined as the equivalent diameter ratio derived from the segment open-area ratio per ISO 5167-6:2022 §3.3 (β² = A_open/A_pipe), not simply √(h/D). Using the correct area-ratio-based β definition is essential — using β = h/D or even β = √(h/D) is the single most common wedge-meter sizing error. The wedge forces a single, stable separation at the apex that is largely independent of Reynolds number below approximately ReD = 10⁵, which makes the wedge meter the best choice for low-Reynolds-number, high-viscosity, and non-Newtonian fluids. Key constraints: h/D range 0.2–0.6 (equivalent to β ≈ 0.377–0.791); ReD ≥ 1×10⁴. The wedge apex angle is typically 60° or 90°; the Cd equation depends on apex angle and h/D. The discharge coefficient from ISO 5167-6 is: C = 0.77 − 0.098 × β^0.8. This is the ISO-standardized formula. Some older datasheets from EMCO (the original wedge-meter patent holder) reference a linear C = 0.77 − 0.09β; this is not the ISO 5167-6 equation and may differ by 1–2% in Cd at high β values.

Wedge meters are inherently asymmetric: the opening can be located at the top (for liquids with entrained solids that settle), at the bottom (for gases with condensate), or at the side (for horizontal liquid flows where sediment is not a concern). This makes them uniquely suited to dirty, slurry, and multiphase-leaning services where an orifice plate's sharp edge would wear and a Venturi's throat would accumulate deposits. The tradeoff is higher Cd uncertainty: the standardized Cd carries ±4% uncertainty (k=2) without calibration, reflecting the complex three-dimensional flow field. For custody transfer, individual calibration is mandatory; for process control and internal accounting, the uncalibrated Cd is normally acceptable. Unlike all other ISO 5167 devices, the wedge meter has no expansibility factor specified in the 2022 edition — it is treated as a liquid-only standard. Using a wedge for gas requires manufacturer calibration that includes expansibility characterization.

Step 1 — Reynolds number: Calculate ReD at minimum flow. If ReD < 5,000, ISO 5167 does not apply — you need a calibrated non-standard device. If the fluid contains solids or is viscous and ReD ≥ 1×10⁴, select a wedge meter per Part 6. If ReD ≥ 5,000 and the fluid is clean, proceed to Step 2. Step 2 — Straight run: Measure available upstream straight run. If less than 10D and pipe modifications are impossible, select a cone meter per Part 5 (validated 3D–5D capability) or a balanced multi-hole meter (not ISO-standardized; requires manufacturer calibration). If 10D or more is available, proceed to Step 3. Step 3 — Pressure loss budget: Calculate allowable permanent pressure loss. If loss budget is tight (e.g., large line, continuous operation, high electricity cost), select a Venturi tube per Part 4 for best recovery. If loss budget is generous, proceed to Step 4. Step 4 — Temperature and erosion: If temperature > 450°C or the fluid carries erosive droplets (wet steam, catalyst fines), select a nozzle per Part 3. If temperature ≤ 450°C and the fluid is clean, proceed to Step 5. Step 5 — Custody transfer or not: If the measurement is for custody transfer and you require standardized traceability without individual calibration, select an orifice plate per Part 2 (gas/liquid) or a nozzle per Part 3 (high-temperature steam). If custody transfer is not required, the choice is economic: Venturi (Part 4) for energy savings, orifice (Part 2) for minimum capital cost, cone (Part 5) for space-constrained installations.

This decision tree reduces a complex multi-variable problem to five sequential checks. The most common mistake is skipping Step 1 — an engineer selects an orifice plate from Part 2, sizes it at normal flow where ReD = 50,000, and only discovers during commissioning that minimum flow produces ReD = 3,000, which is below the β = 0.60 minimum. The meter is non-compliant on day one. Always check Reynolds number at minimum, normal, and maximum flow, and verify against the standard's ReD table before finalizing the β.

The 2022 edition brought several substantive changes beyond the editorial consolidation of all six parts. (1) Part 1 now references the 2022 edition as the sole normative reference — citing ISO 5167-1:2003 in a project specification is technically obsolete and may be rejected during design review. (2) Straight-run requirements were tightened for certain upstream fitting combinations, particularly two elbows in perpendicular planes at high β, based on new CFD and experimental data. A meter sized under the 2003 edition may not comply with the 2022 requirements for the same disturbance. (3) Reynolds-number minimums were reviewed and adjusted; some increased, reflecting better understanding of the transitional regime. (4) Uncertainty statements were revised to use consistent k=2 coverage throughout all six parts. (5) Part 5 (cone meters) was reorganized with clearer calibration requirements and extended validation data. (6) Part 6 (wedge meters) introduced the current Cd equation: C = 0.77 − 0.098 × β^0.8, replacing some legacy manufacturer equations that predated standardization. (7) A new Annex in Part 1 provides guidance on computational fluid dynamics (CFD) as a supplementary tool for evaluating installation effects, though CFD does not replace physical compliance with straight-run requirements.

The practical impact: if your plant was built before 2022 and the original meter sizing used the 2003 edition, a re-validation against 2022 may reveal non-compliance — particularly for high-β orifices with complex upstream piping. This does not mean the meter is unsafe or inaccurate, but it may mean the uncertainty budget needs updating and the commercial or regulatory agreement should reflect the edition change. For new projects, specifying "ISO 5167:2022" without a date suffix is acceptable during procurement, but the final documentation should confirm compliance with the 2022 edition.

(1) Using the wrong β formula: the most frequent is using β = d/D for a cone meter instead of β = √(1 − dc²/D²), or using β = h/D or β = √(h/D) for a wedge instead of the segment open-area ratio per ISO 5167-6:2022 §3.3. Both produce a correct-looking number — the error is subtle because it affects the Cd and expansibility calculations downstream, not the β value in isolation. (2) Applying the orifice-plate expansibility equation to a nozzle or Venturi: the orifice ε is empirical and device-specific. Nozzles and Venturi tubes use the thermodynamic isentropic formulation. Mixing them produces errors that grow with ΔP/P. (3) Ignoring ReD minimums at turndown: the standard's ReD limits apply at all flow conditions, not only at normal flow. If your minimum flow drops ReD below the limit, the Cd correlation is invalid regardless of compliance at normal flow. (4) Specifying a flow conditioner without checking compliance: ISO 5167-2 Annex B defines specific hole patterns, thicknesses, and locations. A generic tube-bundle conditioner from a piping catalog is not necessarily compliant. (5) Measuring pipe diameter once at ambient temperature and assuming it is correct at operating temperature: thermal expansion changes both D and d. At 400°C, a carbon steel pipe's diameter increases by approximately 0.5%, which propagates to approximately 1% on flow if uncorrected.

Part 1 (General): read this first. It governs all five device types. Part 2 (Orifice): the workhorse. β = d/D, 0.10–0.75, ReD: β-dependent, ranging from 4,000 (low β) to 160,000 (high β), Cd from Reader-Harris/Gallagher, empirical ε. Use for 80% of standard applications. Part 3 (Nozzle): for high temperature and high velocity. Thermodynamic ε, ISA 1932 or long-radius. Use for steam above 450°C and high-ΔP gas. Part 4 (Venturi): for energy economics. Cd ≈ 0.99, best pressure recovery, largest physical size. Use when energy cost exceeds capital cost. Part 5 (Cone): for short straight runs. β = √(1 − dc²/D²), 3D–5D capability, calibration recommended. Use in compact installations. Part 6 (Wedge): for low ReD, dirty fluids, and slurries. β from segment area ratio (ISO 5167-6 §3.3), ReD ≥ 1×10⁴, Cd = 0.77 − 0.098β^0.8, asymmetric opening. Use when nothing else works. Before selecting any device, run the five-step decision tree: ReD → straight run → pressure loss → temperature → custody transfer. Answer those five questions, and ISO 5167 will lead you to the right part. Need project-specific guidance? Send your process data through our Contact page — we will return a preliminary part selection and sizing review at no cost.