Flow meter calibration: what the reading depends on besides flow

Flow is among the hardest parameters to calibrate meaningfully, because for most meter types the calibration is only valid for a fluid, a pipe and an installation resembling the one it was performed in.

What the meter is actually measuring

The practical point is that almost none of these measure flow directly. They measure a consequence of flow and convert it using assumptions about the fluid — so calibrating with water and using on oil, or calibrating at one temperature and using at another, changes the relationship the calibration established.

Installation effects

A meter assumes a developed, symmetric velocity profile. Pipework upstream disturbs it: a bend introduces swirl, a valve produces asymmetry, a reducer changes the profile shape. Manufacturers specify minimum straight lengths upstream and downstream for this reason, commonly expressed in pipe diameters, and installing inside those lengths introduces an error that no calibration certificate mentions because the certificate describes a different installation.

This is the central limitation of bench calibration for flow. A meter calibrated on a laboratory rig with perfect straight runs, then installed two diameters downstream of an elbow, is operating outside the conditions its certificate describes. Where accuracy matters, in-situ calibration or a flow conditioner is the answer, not a better bench certificate.

Volumetric and mass flow

A volume of gas depends on its temperature and pressure; a mass of gas does not. So a volumetric flow figure is incomplete without the conditions it refers to, which is why gas flows are so often quoted at normal or standard conditions — and why two figures quoted at different reference conditions are not comparable even though both are in cubic metres per hour.

Coriolis meters measure mass directly and sidestep the issue, which is part of why they are chosen where the measurement matters despite the cost. For everything else, reporting a volumetric flow means reporting the temperature and pressure it applies to.

Turndown and the low end

In-situ or on a rig?

Because installation is so much of the result, flow has an unusually strong case for calibrating the meter where it lives. A rig calibration gives a better number under conditions you do not have; an in-situ comparison against a reference meter in series, or against a weighed or timed collection where the process allows it, gives a worse number under the conditions you actually run. For anything where the flow figure carries commercial or regulatory weight, the second is usually the more defensible evidence, and the two are not mutually exclusive — a rig calibration establishes the meter, an in-situ check establishes the installation.

Where in-situ work is impossible, the honest alternative is to document the installation against the manufacturer's straight-length requirement and to treat any shortfall as an uncertainty contribution rather than pretending it away. An estimated contribution written down beats an unquantified error left out.

Turndown and the low end

Flow meters have a usable range expressed as a turndown ratio, and accuracy deteriorates at the bottom of it — differential pressure meters especially, because the pressure drop varies with the square of flow, so at 10 % of full flow the signal is 1 % of full scale. A meter specified as a percentage of full scale rather than of reading behaves badly at low flow for the same reason a multimeter does at the bottom of a range. Calibrate across the range you actually use, including the lowest flow you intend to trust.