Chasing a 6% flow discrepancy across two magmeters
Two meters on the same line disagreed by six percent. Both were reading correctly.
The plant had two electromagnetic flow meters on what was, on the P&ID, the same line. They disagreed by roughly six percent and had done for months. Both had been calibrated. Both passed their verification routines.
The meters were fine
An electromagnetic meter measures average velocity across its cross-section and infers volume from the pipe bore. It assumes a fully developed, symmetric flow profile. Give it a distorted profile and it will report a number that is internally consistent, repeatable, verifiable — and wrong.
The downstream meter sat three diameters after a partially closed valve. Three diameters is nowhere near enough to re-establish a profile after a flow disturbance of that kind. It was measuring exactly what was in front of it, which was not representative of the line.
What we did
- Relocated the downstream meter to a straight run with five diameters upstream and three downstream
- Confirmed the upstream meter's own straight-run compliance while we were there
- Documented the minimum straight-run requirement on the P&ID so the next modification does not repeat it
The two meters now agree inside one percent. Nothing was recalibrated. The lesson is that a verification routine tells you the meter is healthy; it tells you nothing about whether the meter is installed somewhere it can do its job.
