Flow Calibration: Flowmeters, Samplers and Reference Conditions

Matching units are not enough to compare two flow rates: the fluid, reporting basis and stated conditions must also match. A value in L/min at actual conditions cannot be compared directly with a value referred to a specified temperature and pressure. Calibration must identify what the instrument measures and the relationship established; simply adjusting it to the desired reading does not preserve that evidence.

1. Define flow, fluid and measurement chain

Volumetric flow rate is volume per unit time; mass flow rate is mass per unit time. Their relationship is q_m = ρQ, where density ρ refers to the fluid and conditions of volume flow Q. Specify gas or liquid, composition, range and operating mode. A unit such as kg/h expresses mass; L/min expresses volume, even when the commercial device is described as a “mass” instrument.

For gases, a sensor may determine mass flow and display it as equivalent volume at its configured reference conditions. For liquids, temperature and fluid properties remain relevant: the gas conversion in the example is not a universal model for liquids. Viscosity, density and measurement principle may restrict transfer of a calibration to another fluid.

Identify sensor, indicator, acquisition channel, settings, accessories and configuration version. Distinguish the pump command, flow indication and recorded value. A display test or electrical simulation does not necessarily include physical flow measurement.

2. State the conditions explicitly

For actual flow, state where and at which temperature and pressure the volume is defined. For reference flow, declare temperature, absolute pressure, composition and dry or wet basis. “Standard”, “normalised”, sccm or SLPM without these details do not remove ambiguity: do not automatically assign the same temperature to different instruments.

For gas of unchanged composition, with no leakage or accumulation between the sections and under steady conditions, conservation of mass gives Q_r = Q_a(p_a/p_r)(T_r/T_a)(Z_r/Z_a). Subscripts a and r denote actual and reference conditions; Z is the compressibility factor. Use absolute pressures in the same unit and temperatures in kelvin. Do not insert °C or gauge pressures directly.

Assuming Z = 1 requires justification against the uncertainty needed. If composition or water content changes, this ratio is insufficient without an appropriate model. Use the relevant gas temperature, not automatically the room temperature. Also document whether the instrument already applies a conversion: applying it again may introduce an error.

3. Comparability table

This original table helps prepare the service and review apparently inconsistent readings. Complete it before calculating a deviation.

Select a reference suitable for the range and fluid, with uncertainty useful for the decision. Systems may use volume and time, mass collection, pressure-volume-temperature-time, or comparison with transfer standards. The principle's name alone does not assign an uncertainty: implementation, conditions, corrections and stability of the reference in use matter.

4. Connection and data acquisition

Assess leak tightness and the complete path. Adding a reference in series may introduce a restriction and change pressure or pump performance. In samplers, accessories and collection media affect pressure loss; a different test configuration requires justification. Keep a diagram showing flow direction and sensor locations, without assuming equal volumetric flows at sections under different pressures.

Use compatible media and connections within component limits, with authorised isolation. Do not disconnect pressurised lines or introduce gases or contaminants outside the intended configuration. Prevent contamination of both reference and instrument; any filters must be compatible with the method and assessed pressure loss.

Define stabilisation, points and repetitions according to technology, use and uncertainty target. Record original values, conditions and times over comparable intervals. A stable display plateau does not prove absence of pulsation; a long average may conceal relevant changes. Where applicable, verify signal scaling, volume integration, timer, start/stop and exported data.

5. Results and decision

Calculate e = indication − reference after bringing both values to the same basis; the corresponding correction is c = −e. Report points, units and conditions. Do not extend a factor obtained at one point across the whole range without evidence; comparing against the setpoint does not establish actual flow.

The budget may include the reference in use, repeatability, resolution, stability, leaks, pressure, temperature, density or gas model and timing. Avoid counting contributions already included twice, and consider correlations. For integrated volume, also assess duration, synchronisation and changes in flow: flowmeter uncertainty alone may not describe the complete result.

Distinguish standard uncertainty u from expanded uncertainty U = ku, stating k. Conformity requires a criterion and a decision rule agreed before the test. Keep initial results, authorised adjustments and subsequent results separate; do not reconstruct the initial state from the final one.

6. Simulated example: two readings on different bases

A reference indicates 100.00 L/min of dry gas at actual conditions of 298.15 K and 95.000 kPa absolute. The device instead expresses volume at 293.15 K and 101.325 kPa absolute. Assume, only for this example, identical composition, steady conditions, no leaks and Z_r/Z_a = 1.

The conversion gives 100.00 × (95.000/101.325) × (293.15/298.15) = 92.1854 L/min, rounded to 92.19 L/min. Directly comparing 100.00 and 92.19 would suggest a deviation of approximately 7.81 L/min caused by the different reporting basis. This calculation explains the difference; it does not demonstrate conformity because uncertainty and the applicable criterion are missing.

Over 20.00 minutes, assuming constant flow and disregarding uncertainty only for this illustration, the volume is 2,000.0 L at actual conditions or approximately 1,843.7 L at reference conditions. Labelling both simply “sampled litres” would make a concentration denominator ambiguous. If flow varies, a justified estimate of its time integral is needed.

A flow check does not demonstrate physical collection efficiency, biological recovery or correct particle counting. Link the report, configuration and volume to the use decision. Explore metrological traceability, uncertainty and pressure calibration through the metrology hub.

7. Sources and scope

Checked: 1 October 2026. Table and example are original GuideGxP material. Conditions and programmes in individual publications are not universal GMP requirements.

1. Wright, Nakao, Johnson and Moldover, Gas flow standards and their uncertainty, Metrologia 60 (2023), 015002 — measurement models and contributions; full text consulted.
2. NIST SP 250-80, November 2009 — technical publication on the WGFS system and its stated conditions, not a general standard.
3. INRS MétroPol, Principe général et mise en œuvre pratique du prélèvement, version 3.1, January 2026 — flow verification in occupational hygiene; its tolerances or intervals do not transfer to GMP laboratories.