Coriolis meters have become the default choice for custody transfer and batch charging in chemical processing for a good reason: they measure mass flow and fluid density directly from the tube's vibrational response, without depending on a separate density input or an assumed fluid model the way many volumetric technologies do. That directness is genuinely valuable, and it's also where a quiet assumption tends to creep in — that a single density accuracy specification, verified once against a reference fluid, describes the meter's performance across every product it will ever measure.
It doesn't, because a Coriolis meter's density measurement is a function of the tube's mechanical response to the specific fluid inside it, and that response is calibrated against a limited set of reference fluids spanning a certain density and viscosity range. A specialty chemical blend, a viscous polymer intermediate, or a product carrying entrained gas from an upstream process step can sit outside the conditions the original calibration actually characterized well — and the meter keeps reporting a number with the same apparent precision regardless of whether the fluid in the tube still resembles what it was calibrated against.
Entrained gas is the sharpest version of this problem. Even a small, consistently present gas fraction changes the tube's vibrational damping in a way that biases the reported density and, through it, the mass flow calculation — and unlike a single contaminating event, a process-driven gas fraction shows up on every batch, which makes it a systematic bias rather than a noise term that averages out over a shift's production.
Multi-component blends compound the issue in a different way: density alone doesn't uniquely determine composition once more than two components are involved, so a Coriolis-based concentration inference that works cleanly for a two-component mixture can carry meaningfully wider uncertainty once a third or fourth component enters the blend, even though the meter's raw density reading is unchanged. The uncertainty isn't in the meter at that point — it's in the inference step built on top of it.
Building a defensible uncertainty budget for Coriolis custody transfer means treating the density measurement's accuracy as conditional on the actual fluid in the tube, not as a flat number carried over from the calibration certificate — which requires GUM propagation combined with thermodynamic modeling of how the specific blend's density and phase behavior shift across its real composition and process-condition range, entrained gas included.
That's the difference between a Coriolis meter that looks precise on a datasheet and a custody transfer number that holds up when a counterparty asks where the uncertainty actually comes from — the meter didn't change, but the honest answer to that question depends entirely on which fluid, and which conditions, the budget was actually built around.